Vitamin K3 and vitamin C alone or in combination induced apoptosis in leukemia cells by a similar oxidative stress signalling mechanism

Background Secondary therapy-related acute lymphoblastic leukemia might emerge following chemotherapy and/or radiotherapy for primary malignancies. Therefore, other alternatives should be pursued to treat leukemia. Results It is shown that vitamin K3- or vitamin C- induced apoptosis in leukemia cells by oxidative stress mechanism involving superoxide anion radical and hydrogen peroxide generation, activation of NF-κB, p53, c-Jun, protease caspase-3 activation and mitochondria depolarization leading to nuclei fragmentation. Cell death was more prominent when Jurkat and K562 cells are exposed to VC and VK3 in a ratio 1000:1 (10 mM: 10 μM) or 100:1 (300 μM: 3 μM), respectively. Conclusion We provide for the first time in vitro evidence supporting a causative role for oxidative stress in VK3- and VC-induced apoptosis in Jurkat and K562 cells in a domino-like mechanism. Altogether these data suggest that VK3 and VC should be useful in the treatment of leukemia.


Background
Leukemia is a cancer of the bone marrow and blood. The cause of this hematological disorder is currently still unknown. Today, first-line therapy for the treatment of leukemia includes chemotherapy and radiotherapy. Unfortunately, secondary therapy-related acute lymphoblastic leukemia might emerge following chemotherapy and/or radiotherapy for primary malignancies [1]. Therefore, other therapeutic alternatives based on the reactivation of the apoptotic program should be pursued to eliminate cancer cells [2]. Accordingly, the use of vitamin K3 (VK3, also known as menadione [3]) and vitamin C (VC, also known as sodium ascorbate [4]) alone or in combination (VK3: VC [5]) is highly promising in cancer treatment. Yet, the precise pathway(s) by which VK3 and/or VC induce leukemia cell death are not well established. Moreover, given the complexity of death pathways within a cell, placing these pathways in the proper relationship to the drug trigger is challenging.
During the last three decades, vitamin K3 has been known to display anti-tumor action both in vivo and in vitro in human cancer cell lines [4]. Several observations suggest that vitamin K3 might induce apoptosis -a type of cell death- [6] through different biochemical routes including severe depletion of glutathione and sulfhydrylcontaining proteins and alteration of intracellular Ca 2+ homeostasis [7], activation of c-Jun NH 2 -terminal kinase (JNK, [8]), activation of Fas/Fas ligand system independently of the pro-apoptotic p53 protein [9], activation of Fas-dependent and Fas-independent pathways [10] and NF-B activation [11]. Because most studies have looked at a given pathway in isolation using different cell types, the potential interaction between pathways have often not been addressed. Therefore, the complete mechanism (s) of cell death signalization induced by vitamin K3 in a single cell model remains unclear.
Vitamin C is a water-soluble vitamin effective as antioxidant compound under normal conditions [12]. However, Chen and collaborators [13,14] have shown that pharmacological concentrations of (5-15 mM) vitamin C was prooxidant, generating H 2 O 2 -dependent cytotoxicity toward a variety of cancer cells in vitro and in vivo without adversely affecting normal cells. Interestingly, it has been shown that (10 μM) VK3 in combination of (2 mM) VC kill leukemia cells (e.g., K562 cells) by oxidative stress independent of caspase-3, with a minor percentage of cell displaying mitochondrial depolarisation and DNA fragmentation without chromatin condensation consistent with a necrosis-like cell death [15]. Yet, the mechanism by which vitamin C-induced apoptosis in Jurkat and K562 cells is not yet fully established.
Since it is known that the transcription factor NF-B [16], p53 [17], c-Jun [18] and caspase-3 [19] are involved in apoptosis signaling, we hypothesized that VK3 and VC might induce cell death in leukemia cells through activation of such factors by oxidative stress. To test this assumption, we sought to investigate the molecular mechanism by which VC, VK3 alone or in combination (ratio 100:1 [20]) induce cell death in Jurkat (clone E06-1) and K562 leukemia cell lines and lymphocyte cells in relation to the aforementioned pro-apoptotic transcription factors and caspase-3. Although, the role of vitamin C as therapeutic compound is still controversial [21], understanding the mechanism of vitamins alone or together may provide insight into more effective anticancer therapy.

Results
Vitamin K3 (VK3) and vitamin C (VC) alone or in combination induce apoptosis in Jurkat and K562 cells through anion superoxide radical (O 2 .-)/H 2 O 2 generation and mitochondrial damage Figure 1A and 1B show that VK3 and VC provoked typical morphological features of apoptosis in a concentration-dependent fashion according to the conventional AO/EB staining technique [22]. However, (early/late) apoptosis and necrosis (~20%-50%) morphology of cell death were observed at high VK3 (e.g., 35 μM for both cells) and VC (e.g., 20 mM for Jurkat and > 1 mM for K562) concentrations (insets). Remarkably, K562 cells were the most sensible to VC, whereas both cell lines were more sensible to VK3 than lymphocytes (e.g. Jurkat and K562 cells displayed~40% &~25% AO/EB positive nuclei whereas lymphocytes displayed about 3% AO/EB positive nuclei when exposed to 10 μM VK3). These results prompted us to fully examine the effect of VK3 and VC on Jurkat cell line. Figure 1C shows that VK3 produced both anion superoxide radical (O 2 .-) and H 2 O 2 in a concentration-dependent fashion up-to 20 μM, e.g. 94% and 43% percent of formazan ( Figure 2A) and DCF positive ( Figure 2B), respectively in Jurkat cells. Figure 1C also shows that increasing concentrations of VK3 induced mitochondrial depolarization, as reflected by DiOC 6 (3) non-fluorescent cells ( Figure 2C) in a concentration-dependent fashion. Interestingly, apoptotic morphological alterations ( Figure 2D) were associated with mitochondrial depolarization. Since high VK3 concentrations (e.g. > 35 μM) induced nuclear and cytoplasmic destruction of cells (necrosis), thus avoiding a clear-cut evaluation of apoptotic morphology (data not shown), 10 μM VK3 was selected for further experiments.
Recently, it has been shown that mM concentrations of ascorbate were cytotoxic to neuroblastoma cells mediated by H 2 O 2 [23]. Figure 1D shows that VC was capable to generate either O 2 .or H 2 O 2 at the indicated concentrations. Moreover, VC was moderately or highly toxic to Jurkat and K562 cells, respectively at high concentrations. Of note, whilst 1 mM VC induced apoptotic/necrotic morphology in K562, 10 mM concentration induces apoptosis in Jurkat to a similar extend provoked by (10 μM) VK3. Interestingly, VC was innocuous for lymphocyte cells at the indicated concentrations.
VK3 and VC alone or in combination induce apoptosis through of reactive oxygen species CP55,940 [24] and N-acetyl-cysteine (NAC) antioxidants dramatically reduced apoptosis to almost control values and importantly abrogated H 2 O 2 when cells where treated with VK3, VC alone or VK3 in combination with VC ( Table 2). A similar protective effect was observed when K562 cells were exposed to (20 μM) VK3, (0.5 mM) VC and VC: VK3 in a ratio 300:3 and antioxidants (data not shown).

Discussion
In the present study, we provide for the first time in vitro evidence supporting a causative role for oxidative stress in VK3-, and VC-induced apoptosis in Jurkat and K562 cells in a domino-like mechanism involving O 2 .-/ H 2 O 2 , mitochondrial depolarization, transcription factor activation such as NF-, p53, and c-Jun converging in caspase-3 activation and apoptotic morphology. Most importantly, it is shown that high concentration of VC (e.g. 10 mM) alone or in combination with VK3 (e.g. 10 μM) in a ratio 1000:1 and 100:1 induced apoptosis in Jurkat and K562 cells, respectively by a comparable mechanism to VK3 (e.g. 10 μM) and VC (e.g. 10 mM) alone. Moreover, by using antioxidant compounds, we demonstrated that O 2 .-/H 2 O 2 production are essential in VK3-and VC-induced cytotoxicity. This notion was further reinforced by the fact that the antioxidant CP55,940 and NAC completely protect leukemia cells against VK3-and VC-superoxide radicals and H 2 O 2 toxicity. Most importantly, lymphocytes were more resistant to cell death induced either by VC or VK3 treatment alone than leukemia cells. We speculate that differences in cell cycle (i.e., lymphocytes are in G o cell cycle and cancer cells are constantly dividing cells), differences in glutathione content and/or differences in gene expression of antioxidant proteins (e.g., catalase, thioredoxin, superoxide dismutase, glutathione peroxidase) may explain resistance and/or vulnerability in lymphocyte and leukemia cells against VK3 and VC exposure. Taken together our results suggest that different compounds with varied chemical properties might present a similar oxidative stress mechanism of action to erode leukemia cells.
We confirm that VK3 [25] and VC at high concentrations [13,14,26] [26]. How H 2 O 2 might be involved in cell death process? Takada and co-workers [27] have demonstrated that H 2 O 2 induces NF-B activation in Jurkat cells through the spleen tyrosine kinase (Syk). Alternatively, H 2 O 2 may activate NF-B through phosphorylation IBα kinase (IKKα) and IKKβ [28]). In accordance with these observations, we found p65-DAB + nuclei by immunohistochemistry technique in both cell lines treated with either VK3 or VC, as manifestation of p65 activation and translocation to the nuclei, but almost undetectable in untreated cells. Moreover, pharmacological inhibition of NF-B with PDTC inhibited to almost control value the apoptotic morphology. Taken together these data suggest that VK3 and VC induce activation and translocation of the NF-B (p65) probably via H 2 O 2 -induced mechanisms. Our results are in agreement with Jones et al. [12]. However, our results differ from Han et al. [29] who showed VC toxicity in HL-60 cells through down-regulation of NF-B activity. Although the nature of this discrepancy is not known, one possible explanation is that HL-60 cells constitutively express NF-B (also called "NF-B positive" cancer cells), which confers cell survival [30]. We conclude that depending on the inducible or constitutive expression NF-B profile in cells, NF-B will be an important factor to determine the appropriate therapeutic strategy against lymphoblastic/myelogenous leukemia. Consequently, NF-B may constitute an important therapeutic target [31]. In accordance with other scientific reports and our present data suggest that activation of NF-B is linked to cell death signaling in Jurkat and K562 cells [16]. Indeed, NF-B transcribes pro-apoptotic genes such as p53. Immunohistochemical detection of p53 may indicate that p53 can be directly up-regulated by NF-B. Moreover, pharmacological inhibition of p53 by PFT completely abolishes VK3 and VC-evoked apoptosis. In contrast to Karczewski et al. [32], our data comply with the notion that p53 is involved in the mechanism of VK3 and VC toxicity in Jurkat and K562 cells [33]. It is noteworthy to mention that p53 transcribes pro-apoptotic genes such as PUMA, Noxa, Bim, Bid, Bax, which are able to permeabilize mitochondria, thus promoting the release of the apoptogenic cytochrome c, which elicits caspase-3 protease activation, leading to nuclear chromatin fragmentation, typical of apoptotic morphology. In fact, caspase-3 NSCI inhibitor protected cells from those noxious stimuli. In contrast to Verrax et al. [15], our data clearly showed that VC, VK3 or VC/VK3 induced toxicity in Jurkat and K562 cells by caspase-3 dependent form of cell death. Jurkat cells (1 × 10 6 cell/mL) were left untreated or treated with VC or VK3 alone or in combination at the indicated concentration for 24 h. After this time, mitochondrial membrane potential and nuclear morphological changes were evaluated using DiOC 6    According to immunohistochemical staining and pharmacological blockade, we found that VK3 and VC activate c-Jun. How, then, these concurrent signaling pathways are activated either by VK3, VC or H 2 O 2 ? As mentioned above, H 2 O 2 can directly activate NF-B through Syk or indirectly by activation of the multisubunit IKKα/β by Syk or MEKK1/MAPKK kinase. Therefore, one possibility is that once MEKK1 is activated, it may serve as a cross talk molecule between the JNK and NF-B pathway [34]. Taken together our findings and this information, it is reasonable to assume that NF-B and JNK/c-Jun are accomplices with each other during VK3/ VC-induced apoptosis in leukemia cells. Furthermore, JNK interacts with p53 in response to stress [35]. We conclude that induction of NF-B, JNK/c-Jun, and p53 by VK3 and VC might be particularly suitable for the treatment of leukemia [31,36].
During the last decade, VC and VK3 administered in a ratio 100:1 respectively, exhibit synergistic anti-tumor activity by a cell death process denominated autoschizis. This cell death is a type of necrosis characterized by exaggerated membrane damage and progressive loss of cytoplasm through a series of self-excisions [37]. Because cell death is a pure morphological phenomenon [6], we used acridine orange/ethidium bromide (AO/EB) staining as one of the most reliable and unbiased method to identify live, (early and late) apoptotic and necrotic cell [22] compared to other standard methods [38]. Therefore, we first consistently found that (10 μM) VK3 and (10 mM) VC induced (early and late) apoptosis judged against N,N,N′,N′-tetrakis-(2-Pyridylmethyl)ethylenediamine (5 μM, TPEN) reagent, which induces 100% nuclei apoptotic morphology in a caspase-3 and p53dependent fashion in Jurkat cell line (control, data not shown). We found a ratio 100:1 VC: VK3 induced no cell death and mitochondrial damage in Jurkat cells. However, a significant percentage of (late) apoptotic morphology and depolarized mitochondria in cells were provoked by a ratio 1000:1 (VC: VK3) compared to cells treated with VC and VK3 alone. In agreement with others data, it is found that a ratio 100:1 VC: VK3 induced cell death and mitochondrial damage in K562. Strikingly, Jurkat and K562 cells showed classical apoptosis and necrosis [6] rather than autoschizis morphology [37]. Furthermore, it was found that (1000:1; 100:1) VC/VK3 activates NF-B, p53, c-Jun and caspase-3, typical of apoptosis. In agreement with Sakagami et al [39] and Ogawa et al. [40], altogether our data suggest that VK3 and VC induce apoptosis in leukemia cells by oxidative stress [41]. In contrast to other reports [15,[42][43][44], our data suggest that autoschizis might not be operative, at least under the present experimental conditions, in Jurkat and K562 cells. Taken together these results imply that VC-and VK3-induced autoschizis is a cell-specific rather than universal cell death process. Altogether our data suggest that a (1000:1; 100:1) VC: VK3 dose should be an intravenously therapeutic dose in the treatment of lymphoblastic and myelogenous leukemia.

Conclusion
We provide evidence that VK3 and VC alone or in combination induces apoptosis in leukemia cells by a sequential cascade of molecular events involving the production of ROS, simultaneous activation of NF-B/ p53/c-Jun transcription factors, mitochondrial depolarization and caspase-3 activation pathway. These data    confirm our hypothesis that VK3 and VC kill leukemia cells by oxidative stress mechanism. Most importantly, VK3 and VC are harmless to lymphocytes, at least under the present in vitro conditions. Taken together our results support the notion that oxidative stress may play an important role in the killing of leukemia cells.

Experiments with leukemia cell line and lymphocyte cells
Morphological assessment of cell death by fluorescence microscopy using acridine orange/ethidium bromide (AO/EB) double staining The cell suspension (1 mL, final volume) was exposed to increasing vitamin K3 (1, 5, 10, 20, 35, 50 μM), vitamin C (ascorbic acid at pH 7; 0.1, 0.5, 1, 5, 10, 20 mM) or 10 mM VC: 10 μM VK3 concentrations freshly prepared in RPMI-1640 medium in the absence or presence of different products of interest for 24 h at 37°C. The cells were then used for fluorescent microscopy analysis as detailed elsewhere [45].

Assessment of superoxide anion radical generation
Superoxide anion radicals were evaluated as described elsewhere [46]. Formazan was quantified under a light microscope (Zeiss Axiostart 50).

Qualitative analysis of mitochondrial membrane potential (Δψ m )
The cancer cell lines were treated as described above and incubated with cationic lipophilic DiOC 6 (3) (1 μM, final concentration) to evaluate Δψ m as detailed elsewhere [45]. Green fluorescent cells (reflecting highpolarized and low-polarized mitochondria) were quantified under a fluorescence microscope.

Photomicrography
The light microscopy and fluorescent photomicrographs shown in figures and supplemental ones were taken using a Zeiss (Axiostart 50) microscope equipped with a Canon PowerShot G5 digital camera.

Statistical Analysis
The aforementioned parameters were quantified by counting a minimum of 500 total cells blind to experimental setting and viewer. The experiments were performed in 3 independent settings. Data are means ± S.D. of three independent experiments. One-way ANOVA analyses with Bonferroni or Games-Howell post-hoc comparison were calculated with SPSS 18 software. A p-value of < 0.05 and < 0.001 was considered significant.