Sevoflurane suppresses neuronal cell viability and induces neuronal cell death
As previously, we reported on the neurotoxic effect of volatile anesthetic sevoflurane on newborn neuronal cells (Piao et al. 2020). HT22 cells from mice are comparable to undifferentiated neuronal cells that are highly vulnerable to oxidative stress. The primary hippocampal neurons were derived from newborn Sprague-Dawley rats. We applied 2%, 4%, and 8% sevoflurane to neuronal cells for 6 h, 12 h and 24 h, respectively, and then used MTT assay to evaluate the cell viability and cell death was assessed by LDH assay. In contrast with the control group, sevoflurane exposure significantly caused a
concentration- and time-dependent reduction in neuronal cell viability (Fig. 1A). Moreover, once neuronal cells exposed to 4% sevoflurane for 6 h, there was an obvious elevation in cell death could be observed, and the ratio of cell death raised continuously with elevated concentration and extended duration of sevoflurane exposure (Fig. 1B). Results of microscopy analysis consistently demonstrated that the majority of neuronal cells, subjected to a 24-h treatment with 8% sevoflurane, shrank in size and became more rounded (Fig. 1C). In a word, these findings indicated that sevoflurane exposure suppressed cellular viability and induced death of neuronal cells.
Sevoflurane triggers neuronal cell ferroptosis
To clarify the potential mechanism underlying sevoflurane-caused neuronal cell death, we further explored whether ferroptosis played a role in neuronal death induced by sevoflurane. Considering that iron-dependent lipid peroxidation is typical feature of ferroptosis (Xie et al. 2016), we determined the effects of sevoflurane on the levels of ferrous iron and MDA (a lipid peroxidation product) in neuronal cells. Compared with control, when neuronal cells were exposed to 4% sevoflurane for 6 h, contents of ferrous iron and MDA have increased significantly, which were further amplified with raised concentration and extended exposure period of sevoflurane (Fig. 2A, B). Therefore, these results indicated that sevoflurane promoted concentration- and time-dependent elevations of ferrous iron and lipid peroxidation in neuronal cells.
To investigate whether intracellular iron could affect lipid peroxidation and resultant cell death, 100 μmol/L iron chelator DFO was administered to neuronal cells an hour prior to sevoflurane exposure, and data showed that suppression of intracellular ferrous iron significantly inhibited MDA production and reduction in cellular viability caused by sevoflurane (Fig. 2C-E). Moreover, present study showed that DFO pretreatment significantly abrogated sevoflurane-triggered neuronal cell death (Fig. 2F). Furthermore, the decreased size of neuronal cells and synapse length caused by sevoflurane exposure was significantly restored in the presence of DFO, as light microscopy observation displayed (Fig. 1C). These findings indicated that increased intracellular iron promoted lipid peroxidation and neuronal death following sevoflurane exposure.
Next, to further verify whether lipid peroxidation accounting for sevoflurane-triggered neuronal cell death, lipophilic antioxidants 50 μmol/L Fer-1 or 10 μmol/L Lip-1 were incubated with neuronal cells for 1 h ahead of sevoflurane exposure. Our results showed that treatment of Fer-1 or Lip-1 distinctly attenuated the elevation of MDA level caused by sevoflurane and prevented the death of neonatal neuronal cells (Fig. 2D-F). In parallel, pretreatment of antioxidants VitE or GSH in neuronal cells obviously attenuated the increase of MDA and restrained the lethal effect of sevoflurane (Fig. 2D-F). These revealed the important role of lipid peroxidation in sevoflurane-triggered neuronal cell death. Moreover, transmission electron microscopy imaging of HT22 cells showed that sevoflurane-treated cells had shrunken mitochondria (M) with expanded cristae, irregular nuclei (N) with multiple nucleoli (Nu), and locally apparent expansion of rough endoplasmic reticulum (RER), elevated membrane density and perinuclear space widened slightly as well (black arrows) (Fig. 2G), which were in accordance with ferroptosis-like morphological features (Tang et al. 2021a). Thus, sevoflurane exposure triggered neonatal neuronal cells undergoing ferroptosis.
To further elucidate the exact mechanism underlying aberrant elevation in intracellular iron induced by sevoflurane, then we detected the changes in iron regulatory proteins (Xie et al. 2016). In the current study, we observed that transferrin receptor (TFR), which is responsible for importing iron into cells, was obviously upregulated in a method that depended on concentration and duration of sevoflurane exposure (Fig. 2H, I). Meanwhile, after sevoflurane exposure, ferroportin (FPN) that is a known efflux iron pump was markedly downregulated in neuronal cells, whereas there has been no significant change in FTH and FTL, which compose ferritin and are in charge of binding divalent iron ions to form iron storage protein complexes (Fig. 2H, J). To sum up, these data revealed that sevoflurane promoted intracellular ferrous iron overload through upregulating TFR and downregulating FPN, not via regulating ferritin.
JNK/p38 MAPK pathways are implicated in sevoflurane-triggered ferroptosis of neuronal cells
Serving as critical stress-responsive pathways, JNK/p38 MAPK signals are activated in response to a variety of cellular stresses, and consequently play a vital role in the pathogenesis of apoptosis, Parthanatos and ferroptosis (Hotamisligil and Davis 2016; Ye et al. 2019; Zhao et al. 2023; Zheng et al. 2017). Thus, we explored whether JNK/p38 MAPK activation was involved in neuronal cell ferroptosis triggered by sevoflurane. Western Blotting was used to test the changes in JNK, p-JNK, p38 and p-p38 in the condition of different exposure duration and concentration of sevoflurane. We found that sevoflurane obviously elevated the levels of p-JNK and p-p38 expression (Fig. 3A-C). On the contrary, suppression of JNK and p38 MAPK signal pathways with specific inhibitor SP600125 and SB203580 obviously alleviated neuronal cell death as well as elevations of ferrous iron and MDA following sevoflurane exposure (Fig. 3D-J). Moreover, we observed that sevoflurane-triggered TFR upregulation and FPN downregulation were both apparently mitigated while JNK/p38 MAPK pathways were blocked (Fig. 3K-M). Thus, these findings suggested that activated JNK/p38 MAPK may critically contribute to sevoflurane-induced neuronal cell ferroptosis.
Activation of JNK/p38 MAPK promotes sevoflurane-induced DNA damage in neuronal cells by increasing intracellular ROS
DNA damage has been widely regarded as a crucial player in cell death which was related to various pathological conditions, including neurological diseases and ischemic injury (Caldecott et al. 2022; Li et al. 2018). Nevertheless, it is not clear whether sevoflurane could induce DNA damage in neonatal neuronal cells. The alkaline comet assay was used to detect both single-strand breaks (SSBs) and double-strand breaks (DSBs), the results of which revealed that sevoflurane-exposed cells had more DNA strand breaks, which exhibited more cells with longer comet tails in comparison with the control cells (Fig. 4A). Similarly, neutral comet assay, which only detects DSBs, also showed that cells treated with sevoflurane encountered more DNA DSBs in a concentration-dependent way (Fig. 4A). Additionally, a specific biomarker of DNA DSBs, phosphorylation of histone variant H2AX at Ser139 (γ-H2AX), was observed by fluorescence microscope, which displayed that sevoflurane triggered more γ-H2AX foci formation in neuronal cells compared to the control group, and these manifestations are consistent with results by western blotting showing that a visible γ-H2AX upregulation caused by sevoflurane (Fig. 4B-D). Therefore, these findings showed that sevoflurane exposure resulted in DNA damage in neuronal cells.
While ROS is a known crucial cause leading to DNA damage (Yuan et al. 2020), the activated JNK/p38 MAPK have been proposed to exacerbate intracellular ROS production, linking the process of DNA damage to the pathogenesis of cell death (Hocsak et al. 2017; Roos et al. 2016; Zheng et al. 2017). However, the role of JNK/p38 MAPK-triggered ROS production in DNA damage induced by sevoflurane in neuronal cells remains controversial. Quantitative analysis of DCFH-DA fluorescence intensity was used to evaluate intracellular ROS level and results suggested that sevoflurane exposure improved intracellular ROS generation in a method depends on concentration (Fig. 4E). Additionally, pretreatment with NAC (5 mmol/L) to scavenge ROS could obviously attenuated sevoflurane-induced DNA damage (Fig. 4A-B, E-G). These indicated that ROS promoted DNA damage caused by sevoflurane in neuronal cells. Meanwhile, we noticed that pretreatment with SP600125 and SB203580 to suppress phosphorylation of JNK and p38 MAPK markedly mitigated increased intracellular ROS and DNA damage triggered by sevoflurane exposure (Fig. 4A-B, J-L), suggesting that activation of JNK/p38 MAPK contributed to sevoflurane-induced ROS production and subsequent DNA damage in neuronal cells. Conversely, suppression of ROS with antioxidant NAC obviously prevented sevoflurane-caused upregulation of p-JNK and p-p38 as well as neuronal cell death (Fig. 4F, H, I, P). Moreover, we observed that exterior H2O2 exposure resulted in phosphorylation of JNK and p38 in neuronal cells (Fig. 4M-O). Thus, we concluded that sevoflurane induced positive feedback between JNK/p38 MAPK activation and ROS generation in neonatal brain, which regulated mutually and led to the excessive ROS accumulation and resultant DNA damage.
ATM/p53 activation accounts for sevoflurane-triggered ferroptosis of neuronal cells
It has been documented that DNA damage caused by ROS could trigger the activation of ATM/p53 pathway (Roos et al. 2016; Srinivas et al. 2019), leading to autophagy, apoptosis, and ferroptosis (Chen et al. 2020; Makarevich et al. 2020; Tian et al. 2021). Thus, we further verified whether activated ATM/p53 involved in neuronal cell ferroptosis induced by sevoflurane. Western blotting analysis showed apparent increases in ATM, p-ATM, p53, and p-p53 levels in a way that was based on concentration and time, while neuronal cells were treated with 4% or 8% sevoflurane for 6 h, 12 h or 24 h (Fig. 5.1A-F). Conversely, administration of ATM inhibitor KU60019 and p53 inhibitor PFT-α could counteract sevoflurane-induced activation of ATM and p53 (Fig. 5.1G-I). Consistently, confocal microscopy observation also confirmed sevoflurane exposure markedly elevated the phosphorylated level of p53 in neuronal cells, which was diminished by KU60019 and PFT-α treatment (Fig. 5.1J). These indicating that sevoflurane exposure activated ATM/p53 pathway in neonatal neuronal cells. Then, to reveal the role of ATM/p53 pathway in neuronal ferroptosis triggered by sevoflurane, pharmacological or genetic inhibition of ATM/p53 pathway was applied in neuronal cells. We observed that KU60019 and PFT-α or siRNA targeting p53 markedly repressed the elevations of intracellular iron and MDA levels as well as cell death caused by sevoflurane exposure (Fig 5.2A-F), suggesting that activated ATM/p53 promoted ferroptosis of the neuronal cells following sevoflurane exposure. Next, we explored why ATM/p53 could promote an aberrant rise in ferrous iron by detecting TFR, FPN, and metal-regulatory transcription factor 1 (MTF1), which accounts for triggering expression of FPN (Troadec et al. 2010), with or without ATM or p53 inhibitors. Results showed that suppressing ATM with KU60019 obviously reversed sevoflurane-caused upregulation of TFR and downregulation of FPN and MTF1 (Fig. 5.2G, H). Meanwhile, PFT-α or p53 siRNA could inhibit the increase in TFR expression induced by sevoflurane (Fig. 5.2G-J). Of note, we also observed that sevoflurane-induced the phosphorylation of ATM and p53 was markedly suppressed when JNK/p38 MAPK pathways were blocked (Fig. 5.2K-M), which revealed the role of activated JNK/p38 MAPK acting as upstream effectors to modulate ATM/p53 activation in sevoflurane-induced ferroptosis of neuronal cells. Overall, our results suggested that ATM/p53 promoted the increase in intracellular iron caused by sevoflurane via upregulating TFR and downregulating FPN, which resulted in iron-dependent lipid peroxidation, ultimately leading to neuronal cell ferroptosis.
Activation of ATM/p53 promotes sevoflurane-induced lipid peroxidation by activating NOX4, ALOX12 and ALOX15 and suppressing SLC7A11
Hydrogen peroxide (H2O2) could interact with liable ferrous iron, classically referred to as the Fenton reaction, and then to produce hydroxyl radicals, which subsequently peroxidize PUFAs to drive lipid peroxidation production, ultimately leading to ferroptosis (Lee and Hyun 2023). NADPH Oxidase 4 (NOX4) is in charge of transferring an electron from NADPH to molecular oxygen to generate superoxide, then transformed into H2O2 in the presence of SOD (Bedard and Krause 2007; Sun et al. 2016). Fluorescence microscopy was conducted to visualize superoxide and representative images revealed that sevoflurane induced an apparent elevation in superoxide level in neuronal cells (Fig. 6.1A), which was consistent with upregulated NOX4 caused by sevoflurane (Fig. 6.1D, E). In contrast, pretreatment with the ATM and p53 inhibitors markedly counteracted the increases in NOX4, superoxide and H2O2 after sevoflurane exposure (Fig. 6.1H-J, A, B), suggesting that activated ATM/p53 triggered NOX4 upregulation and subsequent superoxide and H2O2 production in sevoflurane-treated neuronal cells. As iron-containing enzymes, arachidonate lipoxygenase 12 (ALOX12) and arachidonate lipoxygenase 15 (ALOX15) are responsible for oxygenating polyunsaturated phosphoethanolamines (PE) and catalyzing PUFAs peroxidation to generate lipid-based ROS, which resulting in lipid peroxidation (Kang et al. 2021), thus we tested the changes in ALOX12 and ALOX15 in neuronal cells secondary to sevoflurane exposure. Results revealed that sevoflurane induced apparent upregulation of ALOX12 and ALOX15, which were obviously counteracted by pretreatment with PFT-α (Fig. 6.1D, F-J). Moreover, we found that inhibition of p53 by siRNA-mediated knockdown not only markedly prevented increases in NOX4, superoxide and H2O2 caused by sevoflurane exposure (Fig. 6.1K-M, A, C), but also reversed sevoflurane-induced the upregulation of ALOX12 and ALOX15 (Fig. 6.1K-M). The combined data indicated that ATM/p53 activation promoted lipid peroxidation caused by sevoflurane via upregulating NOX4, ALOX12 and ALOX15.
As a well-known primary antioxidant, glutathione (GSH) that derives from cysteine constitutes an endogenous antioxidant system, which could sense oxidative stress damage and serve to scavenge H2O2 and lipid peroxidation in the presence of GPX4 (Wei et al. 2023). We observed that sevoflurane induced significant decrease of GSH in parallel with reduction in GPX4 expression in neuronal cells (Fig. 6.2J, A, C). While amino acid antiporter system Xc- is known to facilitate the exchange between extracellular cystine and internal glutamate, solute carrier family 7 member 11 (SLC7A11) which encoded light chain of system Xc-, was markedly downregulated after sevoflurane exposure along with a decrease of cysteine (Fig. 6.2A, B, M). Furthermore, the suppression of ATM/p53 pathway could markedly alleviated SLC7A11 downregulation as well as cysteine and GSH depletion induced by sevoflurane (Fig. 6.2D-I, N, O, K, L). Remarkably, pharmacological and genetic inhibition of p53 could not prevent downregulation of GPX4 caused by sevoflurane (Fig. 6.2D-I), suggesting that GPX4 is dispensable for p53-mediated lipid peroxidation. Collectively, these data indicated that ATM/p53 activation promoted sevoflurane-induced lipid peroxidation by improving NOX4, ALOX12 and ALOX15 and repressing SLC7A11, ultimately leading to neuronal cell ferroptosis.
Sevoflurane-triggers neuronal ferroptosis in hippocampi of neonatal mice
The death of neuronal cells and long-term cognitive deficits have been reported previously in neonatal mice after repeated sevoflurane exposure (Liu et al. 2022; Tang et al. 2021c). Thus, newborn mice encountered 3% sevoflurane treatment for three consecutive days from postpartum day 6 (P 6) to P 8 two hours per day, and to identify whether ferroptosis was implicated in the death of neonatal neurons induced by sevoflurane. On the 7th day after final anesthesia, we stained the hippocampus with hematoxylin and eosin (HE), and results showed visible hippocampal neuronal injury or death following sevoflurane exposure, as evidenced by sparse cytoplasm in pink, irregular arrangement of neuronal structure with pyknotic nuclei and cell shrinkage (Fig. 7.1A). Based on statistical analysis, the pyramidal neuron survival rate in hippocampal CA1 region was 78% on d 7 after the last sevoflurane treatment (Fig. 7.1B). These findings confirmed that repeated exposure of sevoflurane to neonatal mice can induce hippocampal neuronal death. Moreover, we also found that repeated sevoflurane exposure markedly elevated the levels of ferrous ions and MDA in hippocampi at 6 h after final anesthesia, and these effects were counteracted by iron chelator DFP (Fig. 7.1C, D). Furthermore, pretreatment with DFP attenuated hippocampal neuronal injury induced by sevoflurane and improved the number of surviving CA1 pyramidal neurons (Fig. 7.1A, B). Collectively, these data supported that repeated sevoflurane exposure triggered neuronal ferroptosis in hippocampi of neonatal mice.
To explore the role of activated ATM/p53 in hippocampal neuronal ferroptosis caused by sevoflurane, we observed changes in ATM/p53 pathway-related proteins. Results showed that repeated exposure to sevoflurane obviously increased the levels of ATM, p-ATM, p53, and p-p53 in hippocampi, indicating sevoflurane activated ATM/p53 pathway (Fig. 7.2A-D). Additionally, suppressing p53 with PFT-α not only obviously repressed death of hippocampal neurons and increases in ferrous ions and MDA caused by sevoflurane (Fig. 7.1A-D), but also reversed sevoflurane-induced TFR increase in hippocampi (Fig. 7.2E, F). Furthermore, we found that sevoflurane exposure apparently upregulated the expression of NOX4, ALOX12 and ALOX15, also downregulated SLC7A11 expression and decreased levels of GSH and cysteine in the hippocampi, and these effects were alleviated by PFT-α treatment (Fig. 7.2E-H). These findings indicated that ATM/p53 activation promoted hippocampal neuronal ferroptosis caused by sevoflurane in neonatal brain. Considering that JNK/p38 MAPK are well acknowledged as important signals to exacerbate intracellular ROS production and ATM/p53 pathway can be activated by DNA damage through ROS-mediated oxidative stress (Plastira et al. 2020; Srinivas et al. 2019), we examined whether ATM/p53-mediated neuronal ferroptosis induced by sevoflurane was related to JNK/p38 MAPK-dependent oxidative DNA damage. We observed overt rises in p-JNK and p-p38 and increased levels of ROS and γ-H2AX in hippocampi upon repeated sevoflurane exposure, which was markedly inhibited by SP600125 and SB203580 pretreatment (Fig. 7.2I-M). Moreover, our results showed that the increased levels of intracellular iron, MDA, p-ATM and p-p53 as well as neuronal death caused by sevoflurane, were all obviously mitigated by SP600125 and SB203580 (Fig. 7.1A-D and 7.2I, J). Furthermore, we also observed that pretreatment with NAC to scavenge ROS could obviously attenuated sevoflurane-induced upregulation of γ-H2AX, phosphorylation of ATM and p53, and neuronal death in hippocampi (Fig. 7.2A-D, M and 7.1A, B). Therefore, on the basis of that activated JNK/p38 MAPK are causal factors for ROS overproduction that led to DNA damage, we concluded that JNK/p38 MAPK acted as upstream regulators of intracellular ROS accumulation in sevoflurane-induced neuronal death, which signaled to induce DNA damage and resultant ATM/p53 activation, leading to neonatal hippocampal neuronal ferroptosis.
Sevoflurane-induced spatial memory deficit is associated with activated ATM/p53-mediated ferroptosis
Given the significance of hippocampus in learning and memory, we assessed the spatial memory abilities of mice encountered repeated sevoflurane exposure by using Morris water maze on postnatal day 31. As training went on, the mice showed a noticeable decrease in latencies to identify the target platform (Fig. 8A, B), suggesting that they were learning from daily practice. In comparison to control group, the escape latencies in mice exposed to sevoflurane were prolonged on d 3, d 4 and d 5 of the spatial training tests, and as swimming tracks on d 6 showed that mice spent shorter time hanging around in the target quadrant (Fig. 8A-C). Conversely, mice pretreated with DFP, SP600125, SB203580, NAC and PFT-α located the submerged platform in shorter time and spent more time in the target quadrant in contrast to the sevoflurane-treated mice (Fig. 8A-C). In conclusion, these data indicated that the activated ATM/p53-mediated ferroptosis contributed to spatial memory deficit induced by sevoflurane in neonatal mice.