Citrate kills tumor cells through activation of apical caspases

Abstract

Most tumor cells exhibit a glycolytic phenotype. Thus, inhibition of glycolysis might be of therapeutic value in antitumor treatment. Among the agents that can suppress glycolysis is citrate, a member of the Krebs cycle and an inhibitor of phosphofructokinase. Here, we show that citrate can trigger cell death in multiple cancer cell lines. The lethal effect of citrate was found to be related to the activation of apical caspases-8 and -2, rather than to the inhibition of cellular energy metabolism. Hence, increasing concentrations of citrate induced characteristic manifestations of apoptosis, such as caspase-3 activation, and poly-ADP-ribose polymerase cleavage, as well as the release of cytochrome c. Apoptosis induction did not involve the receptor-mediated pathway, since the processing of caspase-8 was not attenuated in cells deficient in Fas-associated protein with Death Domain. We propose that the activation of apical caspases by citrate could be explained by its kosmotropic properties. Caspase-8 is activated by proximity-induced dimerization, which might be facilitated by citrate through the stabilization of intermolecular interactions between the proteins.

Introduction

Considering the glycolytic phenotype of most tumor cells, inhibition of glycolysis might become a valuable tool in future anticancer therapy. Indeed, suppression of various steps of glycolysis by 2-deoxyglucose (2DG) [1], 3-bromopyruvate [2], or phloretin [3] or stimulation of mitochondrial activity by dichloroacetate [4], have been shown to overcome drug resistance in rapidly growing tumor cells. Among the agents that can suppress glycolysis is citrate, a member of the Krebs cycle. Citrate is an effective natural inhibitor of glycolysis through the inhibition of phosphofructokinase [5]. However, citrate is also involved in several other cellular processes essential for cell physiology. Thus, citrate is a precursor for the fatty acid biosynthesis in rapidly proliferating cells; inhibition of ATP citrate lyase was shown to suppress tumor cell growth [6]. Further, citrate was found to block pyruvate dehydrogenase activity when administered to rats. This effect was more prominent in skeletal muscle of tumor-bearing animals [7].

Recently, it has been demonstrated that citrate can either kill malignant cells or facilitate cell death induced by conventional antitumor drugs [8]. Oral administration of citrate has also been reported to have an antitumor effect in the clinic, and it was proposed that suppression of glycolysis in tumor cells might be responsible for the beneficial effect of citrate treatment [9, 10]. Hence, it seems important to understand the underlying mechanisms of citrate-mediated cell death in more detail, and how these mechanisms can be utilized in anticancer treatment. Here, we show that alteration of cellular energy metabolism is not the prime cause of citrate-induced tumor cell death which, instead, seems to result from the stimulation of apoptotic pathways by activation of apical caspases.

Materials and methods

Chemicals

Tri-sodium citrate and all other reagents used were of analytical grade and obtained from Sigma-Aldrich (St. Louis, MO, USA). Reagents for protein determination were purchased from Thermo Fisher Scientific (Beverly, MA, USA).

Cells

All cells used in these experiments were cultured in RPMI 1640 complete medium supplemented with 10 % (w/v) heat-inactivated fetal calf serum and penicillin/streptomycin (100 U/ml). For Tet21N cells, 100 μg/ml hygromycin and 200 μg/ml Geneticin were also added to the medium. Cells were grown in a humidified air/CO₂ (5 %) atmosphere at 37 °C and maintained in a logarithmic growth phase for all experiments.

Measurement of Ca²⁺ concentration

The concentration of Ca²⁺ in the medium was monitored with a Ca²⁺-sensitive electrode (Thermo Fisher Scientific).

Western-blot analysis

For Western-blot analysis, cells were collected after treatment, washed with phosphate-buffered saline (PBS), centrifuged at 150 × g for 5 min and the pellet resolved in PBS. For analyzing cytochrome c release, cells were permeabilized with 0.01 % digitonin for 15 min at room temperature and fractionated into supernatant and pellet by centrifugation for 5 min at 16,000 × g. Protein samples were mixed with Laemmli’s loading buffer, boiled for 5 min at 95 °C and subjected to 12 or 15 % sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS–PAGE) at 40 mA, followed by transfer to nitrocellulose membranes for 90 min at 120 V. Membranes were blocked for 1 h with 5 % non-fat dried milk in PBS at room temperature and subsequently probed overnight with the primary antibody of interest diluted in PBS containing 1 % bovine serum albumin, 0.05 % Tween 20, and 0.05 % sodium azide. The membranes were rinsed and incubated with a horseradish peroxidase-conjugated secondary antibody (1:10,000) and visualized by ECL™ (Amersham Biosciences, Buckinghamshire, UK) and X-ray film. The following primary antibodies where used: mouse anti-human cytochrome c, mouse anti-human caspase-2, mouse anti-human FADD, mouse anti-human Bid (1:1,000; BD Bioscience Pharmingen, San Diego, CA, USA), rabbit anti-human GAPDH (1:5,000; Trevigen, Gaithersburg, MD, USA), rabbit anti-human cleaved caspase-3, rabbit anti-human cleaved caspase-9 (1:1,000; Cell Signaling Technology, Danvers, MA, USA), mouse anti-human α-Tubulin (1:1,000; Sigma-Aldrich, St. Louis, MO, USA) and mouse anti-human caspase-8 (1:100; a kind gift of Professor Peter Krammer, Heidelberg, Germany).

Measurement of caspase activity

The measurement of fluorogenic peptide substrate (Peptide Institute, Osaka, Japan) cleavage was performed using a modified version of a fluorometric assay. Cells were pelleted and washed once with PBS. After centrifugation, cells were resuspended in PBS at a concentration of 2 × 10⁶ cells/100 μl; 25 μl of the suspension were added to a microtiter plate and mixed with the appropriate peptide substrate dissolved in a standard reaction buffer (100 mM HEPES, 10 % sucrose, 5 mM DTT, 0.001 % NP-40 and 0.1 % CHAPS, pH 7.25). Cleavage of the fluorogenic peptide substrate was monitored by AMC liberation in a Fluoroscan II plate reader (Labsystems, Stockholm, Sweden) using 355-nm excitation and 460-nm emission wave-lengths.

Morphological assessment of apoptosis

Cells were seeded on glass cover slips incubated with the indicated treatment and after staining with Hoechst (2 μg/ml) for 10 min, examined using a Zeiss LSM 510 META confocal laser scanning microscope.

Determination of glycolytic activity

The effect of citrate on glycolysis was assessed as alteration in lactate production. A total of 1.5 ml of fresh medium was added to cells growing in 6-cm dishes together with various concentrations of citrate or the non-metabolizable glucose analog 2DG. After 6 h of incubation, aliquots of 0.2 ml were taken for lactate determination. The method is based on the enzymatic oxidation of lactate to pyruvate in the presence of NAD as described in Ref. [11]. Lactate was determined spectrophotometrically by measuring the amount of NADH formed at 340 nm.

In addition, the effect of citrate on lactate production was assessed in cytosolic extracts. Cells were harvested, suspended in buffer containing 150 mM KCl, 1 mM MgCl₂, 5 mM Tris, pH 7.4, and 0.025 % digitonin. After 10-min incubation at room temperature, the cytosolic fraction was separated and various concentrations of lactate or 2DG were added to the extracts. Glycolysis was determined at 37 °C by measuring the amount of lactic acid produced after 30 min in a reaction mixture containing 4 mM NaCI, 10 mM MgCl₂, 10 mM KH₂PO₄, 4 mM ATP, 4 mM ADP, 2.8 mM NAD⁺, and 4 mM glucose, pH 7.4, as described in Ref. [11]. The reaction was stopped by 5 % (w/v) trichloroacetic acid. The lactate concentration in the samples was measured as described above.

Results

Citrate-induced apoptosis in various tumor cell lines, including Tet21N neuroblastoma cells, was assessed after 16 h by cytochrome c release (Fig. 1a) and caspase-3-like activity (Fig. 1b). Citrate triggered apoptosis in a concentration-dependent manner (Fig. 1b), with prominent stimulation of caspase-3-like activity starting above 15 mM.

Fig. 1.

Citrate stimulates cell death assessed by cytochrome c release (a) and stimulation of caspase-3-like activity (b) in neuroblastoma Tet21N cells treated with increasing concentrations of citrate for 16 h

Inhibition of phosphofructokinase resulting in suppression of glycolysis and subsequent ATP depletion has been proposed to be the main cause of apoptosis induction by citrate [8]. Indeed, citrate reduced the content of ATP in Tet21N cells (Fig. 2a). However, even 20 mM citrate caused only about 40 % decrease in ATP content; moreover, a similar or even more prominent, reduction in ATP level caused by 2DG had no observable effect on caspase-3-like activity (Fig. 2b) or cell viability (Fig. 2c). In addition, analysis of lactate production in cytosolic extracts in the presence of various concentrations of citrate (Fig. 2d) or assessment of lactate excreted from the cells into the medium (Fig. 2e) revealed that there was only a modest difference in the extent of glycolysis inhibition caused by 10 and 20 mM citrate, suggesting that inhibition of glycolysis could not explain the strong stimulation of cell death observed at 20 mM citrate (Fig. 1b). Thus, it appears that the inhibition of glycolysis and associated reduction in ATP level per se are not responsible for apoptosis stimulation by high concentrations of citrate. A decrease in intracellular free Ca²⁺ has been hypothesized to be another potential cause of citrate-induced cell death, based on the Ca²⁺-chelating property of citrate, and the finding that exogenous Ca²⁺ added to cells could partially protect them from citrate-induced death [12]. In our experiments, citrate did cause a decrease in Ca²⁺ concentration in the medium, as assessed by a Ca²⁺ sensitive electrode (Fig. 2f). The decrease in Ca²⁺ -concentration observed with 20 mM citrate was similar to the reduction caused by 250 μM of the Ca²⁺ chelating agent EGTA. Restoring the Ca²⁺ level in the medium partially attenuated citrate-induced cell death (data not shown). However, in contrast to citrate, EGTA failed to activate caspase-3 (Fig. 2g). Thus, chelation of Ca²⁺ alone is also not enough to kill cells, but might facilitate citrate-induced cell death.

Fig. 2.

Citrate stimulates cell death irrespective of glycolysis inhibition or Ca²⁺ chelation. Comparison of ATP level (a) and caspase-3-like activity (b) in neuroblastoma Tet21N cells treated with increasing concentrations of citrate or 5 mM 2-deoxyglucose (2DG) for 16 h, (c) images of control Tet21N cells and cells treated with 5 mM 2DG for 16 h, assessment of lactate production in cell-free system in the presence of increasing concentrations of citrate and 5 mM 2DG (d), and lactate content in the incubation medium after cell treatment with citrate or 2DG (e), comparison of Ca²⁺ chelation (f), and caspase-3-like activity (g) in neuroblastoma Tet21N cells treated with citrate or 250 μM EGTA for 16 h

Analysis of the sensitivity to citrate-induced apoptosis of a variety of tumor cell lines, such as neuroblastoma cells [Tet21N, SK-N-AS, SK-N-SH, and SK-N-BE(2)] and non-small cell lung cancer cells (U1810), was assessed after 16 h incubation by measuring caspase-3-like activity, and revealed apparent differences between the cell lines (Fig. 3a). Most striking was the resistance to citrate-induced apoptosis by SK-N-BE(2) cells; the other cell lines showed varying degree of apoptosis upon citrate treatment. Of note, there was no difference in citrate-induced cell detachment between Tet21N cells, which are sensitive to citrate, and the resistant SK-N-BE(2) cells (Fig. 3b). This suggests that cell detachment was a consequence of Ca²⁺ chelation by citrate, rather than a cell death manifestation. Assessment of morphological changes typical of apoptotic cell death confirmed the resistance of SK-N-BE(2) cells to citrate; no signs of DNA damage or fragmentation were observed after citrate treatment (Fig. 3c). Neither was there any detectable release of cytochrome c from mitochondria in SK-N-BE(2) cells upon treatment with citrate (Fig. 3d).

Fig. 3.

Induction of apoptosis by citrate in various malignant cell lines. (a) Citrate-induced caspase-3-like activity in various malignant cells. (b) Analysis of cell detachment in sensitive (Tet21N) and resistant [SK-N-BE(2)] cells treated with increasing concentrations of citrate. (c) Cells with nuclear pyknosis caused by 20 mM citrate. (d) Release of cytochrome c from mitochondria in various citrate-treated cells. In all experiments, cells were incubated with citrate for 16 h

The key executioner enzyme of apoptosis, caspase-3, can be activated via the mitochondrial pathway or via direct processing and activation by caspase-8. Since activation of caspase-3 was observed also at low citrate concentrations at which cytochrome c release was marginal or not detectable, it seemed possible that the caspase-8-dependent pathway might be predominant for apoptosis induction by low concentrations of citrate.

Indeed, assessment of caspase-8 processing using Western blot revealed that this initiator caspase was activated in a dose-dependent manner by 5, 10, and 20 mM citrate (Fig. 4a). Interestingly, SK-N-BE(2) cells, which were resistant to citrate-induced apoptosis, demonstrated almost no detectable caspase-8 protein. Similar results were obtained when the activity of caspase-8 was measured (Fig. 4b). In addition to caspase-3 cleavage, active caspase-8 can cleave the pro-apoptotic protein Bid and thereby induce mitochondria-dependent activation of caspase-3 through permeabilization of the OMM and release of cytochrome c. Indeed, as shown for U1810 cells treated with citrate, cleavage of Bid causes release of cytochrome c and caspase-9 processing, which become detectable at 10 mM citrate (Fig. 4c). Apparently, the activation of caspase-3 by citrate might involve the mitochondrial pathway.

Fig. 4.

Involvement of caspase-8 in citrate-induced apoptosis in various tumor cells. Processing (a) and stimulation (b) of caspase-8 in various tumor cells. (c) Cleavage of Bid, release of cytochrome c from mitochondria, and caspase-9 processing in U1810 lung cancer cells. Lower blot, loading control. (d) Stimulation of caspase-3-like activity in wild-type Jurkat cells and cells lacking caspase-8. (e) Processing of caspase-8 and cleavage of PARP in wild-type Jurkat cell, cells lacking caspase-8, and cells lacking FADD. In all experiments, cells were incubated with citrate for 16 h

In order to further investigate the significance of caspase-8 expression for citrate-induced apoptosis, we extended our cell model system and included wild-type Jurkat leukemia cells and Jurkat cells deficient in caspase-8. In wild-type Jurkat cells, after 16 h incubation, citrate induced a concentration-dependent stimulation of apoptosis assessed by caspase-3-like activity, whereas the apoptotic response in cells lacking caspase-8 was markedly suppressed (Fig. 4d). To address the question of whether citrate engages the receptor-mediated pathway, experiments were performed with Jurkat cells deficient in Fas-associated protein with Death Domain (FADD). FADD is an adaptor molecule and part of the death-inducing signaling complex (DISC), bridging the Fas-receptor and other death receptors to caspase-8. Despite the absence of FADD, these cells demonstrated similar sensitivity to citrate as the wild-type Jurkat cells (Fig. 4e). Increasing concentrations of citrate decreased the level of full-length caspase-8 in both the wild-type and mutant Jurkat cells, suggesting a mechanism for caspase-8 activation by citrate, which is independent of the DISC-complex.

The importance of caspase-8 for citrate-induced apoptosis was further confirmed in experiments with siRNA silencing of caspase-8 in Tet21N cells (Fig. 5a). Down-regulation of caspase-8 caused suppression of apoptosis induced by high concentrations of citrate, although some PARP cleavage was still observed in caspase 8^−/− Jurkat cells (Fig. 3e), as well as after silencing of caspase-8 in Tet21N cells (Fig. 5a). Together, these results suggest that in addition to caspase-8, another initiator caspase might be involved in citrate-induced apoptosis. Analysis of caspase-2 revealed that citrate caused concentration-dependent processing of this protein (Fig. 5b) and activation (Fig. 5d) in Tet21N cells, as well as in the wild-type and mutant Jurkat cells used in the experiment (Fig. 5c). Analysis of the time course of citrate-induced activation of both initiator caspases revealed a rapid processing of caspases-2 and -8 observable as early as 2 h after citrate administration (Fig. 5e). This resulted in processing of caspase-3, stimulation of caspase-3-like activity, and cleavage of PARP (Fig. 5e, f).

Fig. 5.

Involvement of caspase-2 in citrate-induced apoptosis in various cancer cells. (a) Silencing caspase-8 attenuates caspase-3 processing and PARP cleavage. (b) Processing of caspase-2 in Tet21N neuroblastoma cells treated with increasing concentrations of citrate. (c) Processing of caspase-2 in wild-type Jurkat cells, cells lacking caspase-8 and cells lacking FADD. (d) Stimulation of caspase-2 activity in Tet21N neuroblastoma cells treated with citrate. (e) Time-dependent processing of caspase-8, caspase-2, caspase-3, cleavage of PARP, and caspase-3-like activity in Tet21N neuroblastoma cells treated with 20 mM citrate (f). In all experiments, cells were incubated with citrate for 16 h

In order to confirm the contribution of caspase-2 to citrate-induced cell death, Jurkat cells deficient in caspase-8 were transfected with either shRNA against caspase-2 or a negative control shRNA (mock) using a virus-mediated approach. Suppression of caspase-2 in these cells further attenuated citrate-induced caspase-3-like activity (Fig. 6a) as well as processing of caspase-3 (Fig. 6b). These results show that both caspase-2 and -8 are important for citrate-induced cell death.

Fig. 6.

Silencing of caspase-2 further suppresses caspase-3-like activity (a) and processing of caspase-3 (b) in Jurkat cells lacking caspase-8. Cells were incubated with 20 mM citrate for 16 h

Discussion

The mechanism(s) by which citrate kill tumor cells have been studied by different research groups. Based on citrate’s ability to chelate calcium, its induction of apoptosis was explained by sequestration of Ca²⁺ in the medium, since the addition of exogenous Ca²⁺ partially reversed the pro-apoptotic activity of citrate [12]. Considering the inhibitory effect of citrate on one of the glycolytic enzymes, phosphofructokinase, depletion of ATP was also suggested to underlie citrate toxicity [13], as well as its synergistic effect in combination with cisplatin [8]. However, the results presented in this study show that neither a decrease in ATP level nor chelation of Ca²⁺ is sufficient to explain the induction of apoptosis by citrate, although both events might contribute to citrate-induced cell death.

Different cells were responding to citrate to different extent. SK-N-BE(2) cells were found to be the most resistant to treatment. Regarding the p53 mutation, which is characteristic for the SK-N-BE(2) cells [14], as a possible cause for the resistance to citrate treatment, it is important to mention that citrate-induced cell death is p53-independent, since U1810 cells bearing a p53 mutation [15] turned out to be citrate-sensitive. In addition, no p53 activation could be observed in p53 wild-type-sensitive cell lines in response to citrate (data not shown).

The data obtained clearly demonstrate that citrate induces apoptosis via activation of two upstream initiator caspases, caspase-8 and -2. Importantly, the activation of caspase-8 by citrate does not involve the receptor-mediated pathway, since the lack of a component of this pathway, FADD, did not affect caspase-8 activation. Apparently, upon treatment with low concentrations of citrate, caspase-8 cleaves and activates caspase-3, while at higher doses active caspase-8 also involves the mitochondrial pathway, associated with the cleavage of Bid, the release of cytochrome c, processing of caspase-9 (Fig. 4c) leading to activation of caspase-3.

It has recently been shown that N-alpha-acetylation of caspase-2 can promote the assembly of caspase-2 activating complex leading to activation of this enzyme [20]. A key cofactor in N-alpha-acetylation is acetyl-CoA. Hence, incubation with citrate might stimulate cytosolic acetyl-CoA production by ATP citrate lyase, with subsequent N-alpha-acetylation of caspase-2. However, analysis of the possible involvement of ATP citrate lyase in citrate-induced apoptosis through acetylation of caspase-2 revealed that inhibition of this enzyme by hydroxycitrate had no effect on citrate-induced apoptosis (data not shown). In fact, the inability of hydroxycitrate to prevent apoptosis was not surprising, since it has been reported that inhibition of ATP citrate lyase can cause tumor cell death [6], and hydroxycitrate itself can cause apoptosis [16, 17].

A generally accepted, so-called “induced proximity model” proposes that initiator caspases are activated by adaptor-mediated clustering of zymogens [18]. In experiments using a cell free system, citrate was shown to cause both dimerization and activation of caspase-8, and these two events were correlated [19, 20]. These results were suggested to be explained by the so-called kosmotropic properties of citrate. Kosmotropes are a class of salts that contribute to the stability and structure of water–water interactions. Kosmotropes cause water molecules to favorably interact, which also stabilizes intermolecular interactions in macromolecules such as proteins. Our experiments suggest that kosmotropic properties of citrate might also be involved in the activation of apical caspases in vivo; similar to caspase-8, caspase-2 can also be stabilized and activated by citrate.

In conclusion, a primary strategic problem in cancer therapy is how to selectively activate apoptosis in transformed cells. The ability of citrate to induce tumor cell death, as characterized in this study, might contribute to the anticancer effect of citrate treatment observed in patients [9, 10]. Caspase activation appears to be the main consequence of citrate treatment, although inhibition of glycolysis and chelation of Ca²⁺ may contribute to citrate-induced cell death. All of these properties of citrate might be important for its ability to compromise cell survival and contribute to tumor elimination.