Phosphorylation-Mediated Control of Stress Responses Induced by Nanosecond Pulsed Electric Fields

2.1. Overview of stress response

The fundamental aspects of cellular stress responses are highly conserved among eukaryotes from yeast to humans. Figure 1 shows a simplified scheme for eukaryotic stress responses. Under normal physiological conditions, cells continuously undergo protein synthesis, and the rate of protein synthesis is primarily regulated at the translation initiation step. When cells sense stress, they rapidly activate signal transduction that involves phosphorylation-mediated control of translation initiation factors. Stress-induced alterations in the phosphorylation status of translation initiation factors reduce translation initiation and thereby result in attenuation of general protein synthesis. Because protein synthesis requires significant amounts of energy and materials, transient suppression of general protein synthesis conserves cellular resources and is thus beneficial for cells under stress. For these reasons, stress-induced suppression of general protein synthesis serves as a mechanism for survival. Once the stress ends, translation capacity is rapidly recovered by dephosphorylation of the translation factors. Intriguingly, persistent activation of stress responses is often associated with the induction of cell death, suggesting that timely induction and attenuation of stress responses are both critical for cell survival [10–12]. Although eukaryotic cells share a fundamental stress response mechanism [13, 14], as described below, human cells possess more intricate stress responses, including at least two distinct pathways involving phosphorylation of multiple proteins.

2.2. Stress response mediated by eIF2α phosphorylation

2.2.1. Stress-induced eIF2α phosphorylation and translational suppression

Translation initiation is a critical rate-limiting step in protein synthesis, and eukaryotic translation initiation factor 2 (eIF2) plays an essential role in this process [15]. eIF2 binds to guanine nucleotides, such as GDP and GTP. During initiation of translation, an eIF2-bound guanine nucleotide needs to cycle between GDP and GTP (see Figure 2A for details of GDP-GTP cycling in translation initiation). eIF2 consists of three subunits, α, β, and γ, and phosphorylation of the α subunit of eIF2 (eIF2α) is induced by various forms of stress [14, 16]. Phosphorylation of eIF2α at serine 51 interferes with GDP-GTP exchange on eIF2 [17] and consequently suppresses translation initiation (Figure 2B). The role of eIF2α phosphorylation in inhibition of translation initiation is highly conserved among eukaryotes, and the site of stress-induced phosphorylation (serine 51) is conserved in yeast and humans [13, 14].

Although initiation of translation of most mRNA species is dependent on their 5′ cap and is profoundly affected by eIF2α phosphorylation, a subset of mRNA species is actively translated under stressed conditions via alternative translation mechanisms that are independent of the cap structure. Approximately 2.5% of total mRNA is estimated to be preferentially translated under stressed conditions [20, 21], permitting synthesis of a subset of proteins that play critical roles in control of the stress response. For example, initiation of translation of ATF4 (activating transcription factor 4) mRNA is increased by ER stress caused by accumulation of unfolded proteins in the ER. The 5′ UTR of ATF4 mRNA contains two small open reading frames (upstream open reading frames [uORFs]), which play a critical role in initiation of translation under stressed conditions. ATF4 mRNA encodes a transcription factor that induces gene expression for adaptive responses. Thus, stress-induced eIF2α phosphorylation results in both suppression of general protein synthesis to conserve cellular resources and elevated translation of specific mRNA species for adaptive responses [22, 23].

In human cells, phosphorylation of eIF2α is induced by a wide variety of exogenous as well as endogenous stresses, such as amino acid deprivation, UV irradiation, and accumulation of unfolded proteins in the ER. Although these stresses are sensed by individual mechanisms, multiple stress-induced events converge on a single reaction, namely, eIF2α phosphorylation. Thus, eIF2α phosphorylation integrates various stress-induced signals. The stress-induced signaling pathway in human and mammalian cells that involves eIF2α phosphorylation and downstream ATF4 induction is referred to as the integrated stress response (Figure 3) [12, 24].

2.3. Stress response mediated by 4E-BP1 phosphorylation

Human cells have an alternative mechanism for stress-induced translational suppression, which involves 4E-binding protein 1 (4E-BP1) [38]. As mentioned above, most mRNA species are translated in a cap-dependent manner. eIF4E binds to the cap structure of mRNA and in turn recruits eIF4G and other translation initiation factors, resulting in formation of an active translation initiation complex on the 5′ end of mRNA. 4E-BP1 serves as a negative regulator of this process. In unstressed conditions, 4E-BP1 is highly phosphorylated at multiple sites, suppressing its inhibitory activity. Under energy deprivation and other stressed conditions, phosphorylation of 4E-BP1 is substantially decreased, and 4E-BP1 competes with eIF4G for binding to eIF4E, thereby inhibiting translation initiation (Figure 4) [38].

Phosphorylation of 4E-BP1 is primarily controlled by mTORC1 (mammalian target of rapamycin complex 1), which is a member of the phosphatidylinositol 3-kinase-related family of kinases [39]. Under normal physiological conditions, mTORC1 is catalytically active and suppresses the inhibitory activity of 4E-BP1 by phosphorylation. The kinase activity of mTORC1 is regulated by several cellular proteins, one of which is AMP-activated protein kinase (AMPK). AMPK senses the cellular energy status and negatively regulates mTORC1 activity [40]. AMPK, mTORC1, and 4E-BP1 constitute a stress-responsive mechanism for translational suppression, independent of eIF2α phosphorylation. Thus, two mechanisms, the eIF2α-mediated integrated stress response and the 4E-BP1-mediated mechanism, function in translational suppression in human cells (Figure 4).

3.1. Effects of pulsed electric fields on living organisms

Pulsed electric fields (PEFs) refer to high-voltage electric pulses, which are milliseconds, microseconds, and nanoseconds in duration. PEFs have different effects on living organisms depending on pulse duration (Figure 5). PEFs with duration of milliseconds to microseconds primarily act on the cell membrane and cause pore formation. These membrane pores are suitable for transfer of macromolecules, such as plasmid DNA and drugs [5, 6]. Therefore, exposure of living cells to these PEFs is called electroporation and is commonly used for DNA transfection [7]. In addition, these PEFs are used for the introduction of antitumor drugs, which is called electrochemotherapy [41, 42]. Because PEFs with duration of milliseconds to microseconds primarily act on the cell membrane, these PEFs often induce cell death via cell membrane damage.

Recent advances in electrical engineering have enabled the generation of high-voltage electric pulses for ultrashort periods in the nanosecond range, which are called nanosecond PEFs (nsPEFs). It has become increasingly evident that nsPEFs have unique biological actions distinct from electroporation. The pulse duration of nsPEFs is too short to generate membrane pores large enough for entry of macromolecules. Thus, nsPEFs are generally unsuitable for DNA transfection and cancer electrochemotherapy. Instead, nsPEFs produce very small membrane pores that allow passage of small molecules, such as ions and water [43–45]. Accordingly, exposure of cultured human cells to nsPEFs causes Ca²⁺ influx and membrane blebbing due to ion imbalance across the cell membrane [46–48].

Whereas millisecond-to-microsecond PEFs primarily act on the cell membrane, previous theoretical studies have strongly suggested that nsPEFs exert their effects on both the cell membrane and intracellular components [49, 50]. In accordance, extensive biochemical analyses have proven that nsPEFs elicit various intracellular responses, as described below.

3.2. Cellular responses to nsPEFs

Recent studies have revealed that nsPEFs elicit different intracellular responses in a manner dependent on nsPEF intensity. Figure 6 summarizes the relationship between nsPEF intensity and intracellular responses in human cells. Relatively weak nsPEFs do not cause morphological changes observable under a microscope, growth retardation, or cell death. However, cells rapidly respond to such stimuli by activating multiple intracellular signal pathways, including MAPK pathways [51, 52] and AMPK pathway [53]. Intracellular signaling is mediated by sequential phosphorylation of proteins in these pathways, leading to expression of downstream genes [51, 52]. When moderate-intensity nsPEFs are used, two independent mechanisms for stress responses are activated [54]. Intense nsPEFs efficiently induce cell death in vitro [8] as well as in vivo [9].

When cultured human cells are exposed to intense nsPEFs, cell-type dependency of cell death modes has been observed. For example, apoptosis is induced in HL-60 and Jurkat cells by intense nsPEFs [8, 55, 56], whereas necrotic cell death is elicited in several cell lines, including U937, K562, and HeLa S3 [56, 57], demonstrating that the cellular context determines the mode of cell death. Induction of necrosis by intense nsPEFs is a Ca²⁺-dependent process [48, 58], while nsPEF-induced apoptosis is largely unaffected by the presence or absence of Ca²⁺ [58].

Because of the ability of nsPEFs to induce efficient cell death, many studies have attempted to apply nsPEFs to cancer therapy and have demonstrated their effectiveness in several mouse models [9, 59, 60]. Furthermore, a human clinical trial of nsPEFs for cancer therapy has started [61]. For these reasons, nsPEFs have attracted much interest and are increasingly recognized as a novel method of cancer therapy.

4.1. eIF2α phosphorylation induced by nsPEFs

Most mRNA species are translated in a cap-dependent manner, and eIF2 plays a critical role in this process. Under stressed conditions, the activity of eIF2 for translation initiation is suppressed by phosphorylation of its α subunit (eIF2α) [14]. eIF2α phosphorylation is considered to be a hallmark of induction of the integrated stress response and can be examined by Western blotting using an antibody specific to phosphorylated eIF2α. When cultured cells are exposed to nsPEFs, eIF2α phosphorylation can be detected, indicating that nsPEFs activate the integrated stress response [54]. nsPEF-induced eIF2α phosphorylation can be detected in all cell lines examined so far, which include HeLa S3, HCT116, Jurkat, and mouse embryonic fibroblasts (MEFs). In HeLa S3 cells, eIF2α phosphorylation is detectable within 1 min after nsPEF exposure, persists at high levels for 30 min, and decreases thereafter [54]. GADD34 is known to play a critical role in recovery from the integrated stress response. Under stressed conditions, GADD34 is positively controlled at the transcriptional and translational levels. Consistently, expression of GADD34 is significantly activated in nsPEF-exposed cells, suggesting that GADD34 is involved in recovery from the nsPEF-induced stress response [54].

As described above, cellular responses to nsPEFs are dependent on nsPEF intensity (Figure 6). Relatively mild nsPEFs activate several signal transduction pathways, such as MAPK pathways, but are insufficient to induce eIF2α phosphorylation. Moderate levels of nsPEF intensity are required for induction of eIF2α phosphorylation. Such nsPEFs also cause retardation in cell proliferation but not cell death [54]. Intense nsPEFs induce eIF2α phosphorylation and cell death. Currently, it remains unknown whether the nsPEF-induced stress response positively affects cell survival or facilitates cell death induction.

4.2. Participation of PERK and GCN2 in nsPEF-induced eIF2α phosphorylation

Human and mammalian cells have four stress-responsive protein kinases for eIF2α phosphorylation. These kinases differentially respond to various forms of stress, and at least one of them is activated for eIF2α phosphorylation [14]. Autophosphorylation is critical for the activation of these kinases and can be analyzed by Western blot analysis using antibodies against phosphorylated forms of these kinases. In nsPEF-exposed cells, PERK and GCN2 are activated, as shown by their autophosphorylation [54]. Experiments using PERK and GCN2 knockout cells suggest that these kinases play mutually compensatory roles in nsPEF-induced eIF2α phosphorylation. MEFs lacking either PERK or GCN2 gene display nsPEF-induced eIF2α phosphorylation comparable to that in wild-type cells. However, double-knockout cells lacking both PERK and GCN2 genes exhibit a significant reduction in nsPEF-induced eIF2α phosphorylation [54]. These observations suggest that PERK and GCN2 perform redundant functions in nsPEF-induced eIF2α phosphorylation.

Because PERK is well known to be activated by ER stress, the observation on the nsPEF-induced PERK activation raised the possibility that nsPEFs cause ER stress. To clarify this point, downstream events in the ER stress pathway were analyzed [54]. The signal pathway induced by ER stress generally leads to alterations of gene expression [62]. Transcription of CHOP (CCAAT-enhancer-binding protein homologous protein) gene is a major downstream event in the ER stress response. In addition, mRNA for XBP1 (X-box-binding protein 1) is known to undergo alternative splicing after ER stress induction [63]. However, quantitative RT-PCR analysis of CHOP and XBP1 mRNAs demonstrated that nsPEF-exposed cells showed neither elevated CHOP expression nor altered XBP1 splicing. Furthermore, UV irradiation has been reported to induce activation of GCN2 and transcription of downstream genes, such as GADD45 (growth arrest and DNA damage-inducible 45) [64, 65], but quantitative RT-PCR analysis showed no substantial changes in these mRNAs in nsPEF-exposed cells. These observations indicate that nsPEFs exert their effects in a manner that is distinct from ER stress or UV irradiation, although nsPEFs activate PERK and GCN2 [54].

4.3. Decreased 4E-BP1 phosphorylation by nsPEFs

In addition to the eIF2α-mediated response, a distinct mechanism involving 4E-BP1 phosphorylation is known to play a critical role in stress-induced translational suppression [38]. Under normal physiological conditions, 4E-BP1 is highly phosphorylated by mTORC1 [39], and hyperphosphorylation suppresses its inhibition of cap-dependent translation initiation. Stressed conditions, particularly energy deprivation, reduce mTORC1 activity and result in decreased 4E-BP1 phosphorylation. 4E-BP1 at low phosphorylation status interferes with assembly of translation factors on the cap structure of mRNAs and thereby suppresses general protein synthesis (Figure 4). In nsPEF-exposed cells, 4E-BP1 phosphorylation is rapidly decreased [54], suggesting that nsPEFs activate stress responses mediated by 4E-BP1 phosphorylation. The decrease in 4E-BP1 phosphorylation is indistinguishable between wild-type cells and PERK/GCN2 double-knockout cells, supporting the idea that nsPEFs activate two independent mechanisms.

The decrease in 4E-BP1 phosphorylation following nsPEF exposure suggests that nsPEFs cause a reduction in the catalytic activity of mTORC1. A previous study demonstrated that AMPK is rapidly activated by nsPEFs [53]. AMPK functions as an energy sensor and is activated by elevated intracellular AMP levels, which are primarily caused by energy deprivation [40]. The catalytic activity of mTORC1 is well known to be negatively regulated by AMPK [39]. A previous study has demonstrated that, concomitant with AMPK activation, nsPEFs induce phosphorylation of AMPK substrates, such as acetyl-CoA carboxylase-2 [53], suggesting that nsPEF-activated AMPK also phosphorylates other substrates, including mTORC1. Although mTORC1 catalytic activity has not been examined in nsPEF-exposed cells yet, the above observations suggest that AMPK downregulates mTORC1, leading to reduced 4E-BP1 phosphorylation in nsPEF-exposed cells (Figure 7).

4.4. Suppression of general protein synthesis by nsPEFs

As described above, nsPEFs cause elevated eIF2α phosphorylation and decrease 4E-BP1 phosphorylation, both of which are known to be involved in suppression of general protein synthesis. To test whether exposure to nsPEFs actually leads to translational suppression, measurement of protein synthesis rates was required. To this end, metabolic labeling of newly synthesized proteins with radioactive amino acids was employed [54]. Figure 8 shows an outline of metabolic labeling using ³⁵S-labeled amino acids. Using this method, suppression of general protein synthesis in nsPEF-exposed cells was demonstrated [54]. After nsPEF exposure, overall protein synthesis quickly decreased, and maximum suppression of protein synthesis was observed at 30 min. Protein synthesis in nsPEF-exposed cells recovered to approximately 80% within 2 h. When cells were treated with UV irradiation, general protein synthesis decreased gradually for several hours. Compared to UV irradiation, nsPEFs cause acute translational suppression, and recovery is more rapid than in UV-irradiated cells.