Suppression of mitochondrial oxygen metabolism mediated by the transcription factor HIF-1 alleviates propofol-induced cell toxicity

A line of studies strongly suggest that the intravenous anesthetic, propofol, suppresses mitochondrial oxygen metabolism. It is also indicated that propofol induces the cell death in a reactive oxygen species (ROS)-dependent manner. Because hypoxia-inducible factor 1 (HIF-1) is a transcription factor which is involved in cellular metabolic reprogramming by modulating gene expressions of enzymes including glycolysis pathway and oxygen utilization of mitochondria, we examined the functional role of HIF-1 activity in propofol-induced cell death. The role of HIF-1 activity on oxygen and energy metabolisms and propofol-induced cell death and caspase activity was examined in renal cell-derived RCC4 cells: RCC4-EV cells which lack von Hippel-Lindau protein (VHL) protein expression and RCC4-VHL cells, which express exogenous VHL, and in neuronal SH-SY5Y cells. It was demonstrated that HIF-1 is involved in suppressing oxygen consumption and facilitating glycolysis in cells and that the resistance to propofol-induced cell death was established in a HIF-1 activation-dependent manner. It was also demonstrated that HIF-1 activation by treatment with HIFα-hydroxylase inhibitors such as n-propyl gallate and dimethyloxaloylglycine, alleviated the toxic effects of propofol. Thus, the resistance to propofol toxicity was conferred by HIF-1 activation by not only genetic deletion of VHL but also exposure to HIFα-hydroxylase inhibitors.


HIF-1 is activated in RCC4-EV cells under 20% O 2 conditions. HIF-1 activation was investigated in
RCC4-VHL and RCC4-EV cells cultured under normoxic (20% O 2 ) and hypoxic (1% O 2 ) conditions. The protein expression levels of HIF-1α and HIF-1β were assayed by immunoblot analysis (Fig. 3a). Consistent with a previous report 10 , HIF-1α was constitutively expressed in RCC4-EV cells, even in the presence of 20% O 2 , at levels comparable to those observed in RCC4-VHL cells cultured in the presence of 1% O 2 . HIF-1β expression was stable in both cell types under both O 2 levels. Consistent with HIF-1α protein expression findings, the mRNA levels of downstream genes including glucose transporter 1 (glut1), lactate dehydrogenase A (ldha) and pyruvate dehydrogenase kinase 1(pdk1) were higher in RCC4-EV cells than in RCC4-VHL cells under 20% O 2 conditions (Fig. 3b).
Next, the RCC4-EV and RCC4-VHL cell growth rates were examined using the MTS assay (Fig. 3c). No significant difference in the cell growth rate of RCC4-VHL and RCC4-EV cells was found. However, RCC4-EV cells had a higher level of ATP, as compared with the RCC4-VHL cells (Fig. 3d).

HIF-1 activation is required for resistance to propofol-induced cell death. Treatment of RCC4-EV
cells with the HIF inhibitor, YC-1, reduced the expression of downstream genes such as glut1, ldha and pdk1 in RCC4-EV cells but not in RCC4-VHL cells (Fig. 4a). The YC-1 treatment also significantly increased caspase 3/7 activation in RCC4-EV cells (Fig. 4b) within 6 h of exposure to 50 µM propofol. In contrast, significant impact of YC-1 on caspase 3/7 activation was not observed in RCC4-VHL cells. Next, RCC4-VHL cells were exposed to the HIFα-hydroxylase inhibitors, nPG (100 µM) and DMOG (100 µM). These treatments increased the expression of glut1, ldha and pdk1 in RCC4-VHL cells (Fig. 4c). The treatments also significantly suppressed the caspase 3/7 activation (Fig. 4d) that was induced by 50 µM propofol in RCC4-VHL cells. These findings indicated that HIF activation was required and sufficient for establishment of cell protection against propofol-induced toxicity.
Oxygen metabolism in RCC4 cells. Next, we investigated oxygen utilization and glycolysis in RCC4-EV and RCC4-VHL cells using the assays measuring OCR by Cell Mito Stress Test (Fig. 5a) and ECAR by Glycolysis Stress Test (Fig. 5b). OCR was reduced and ECAR was increased in RCC4-EV cells. The mitochondrial basal OCR was significantly lower in RCC4-EV cells than in RCC4-VHL cells in the presence of 20% O 2 (Fig. 5c,d). The significant differences were also detected in the maximum respiratory rates, non-mitochondrial respiration, and proton leak observed in the RCC4-EV and RCC4-VHL cells ( Fig. 5e-g). In addition, significant differences in mitochondrial mass between RCC4-VHL cells and RCC4-EV cells was observed (Fig. 5h). Thus, we observed metabolic reprogramming from aerobic to anaerobic glucose metabolism in the RCC4-EV cells.
The local anesthetic lidocaine induced cell death by targeting mitochondria ETC as well as propofol. To investigate the mode of targeting of propofol, we examined OCR, which depends on the activity of mitochondrial respiratory chain complexes I-IV in membrane-permeabilized and intact cells, using an extracellular flux analyzer (Supplementary Figure 1). Propofol suppressed ETC complex I, II and III-dependent OCR but lidocaine suppressed only complex I (Supplementary Figure 1c,

RNA-Seq analysis of RCC4-EV and RCC4-VHL cells.
We conducted a comprehensive gene expression analysis using RNA-Seq (Table S1) because a line of reports demonstrated that HIF-1 determines oxygen utilization and glucose metabolism [14][15][16] . Our RNA-Seq analysis also made it clear that there are differences in the cellular hypoxic pathway and HIF-1 signaling pathway in RCC4-EV and RCC4-VHL cells ( Fig. 6a; Table S2). RNA-seq identified differences in the expression levels of selected genes within GO:0061621 (canonical glycolysis) (Fig. 6b) and GO: 0004740 (pyruvate dehydrogenase (acetyl-transferring) kinase activity) (Fig. 6c) in these cell lines. To confirm the experimental result of gene expression, we performed meta-analysis using FASTQ files deposited in the Sequence Read Archive (https://trace.ddbj.nig.ac.jp/dra/indexe.html) as SRR1554431, SRR1554986, SRR1554988 and SRR155499. Comparative analysis of gene expression differences between RCC4-EV cells and RCC-VHL cells demonstrated that our study was consistent with the data in SRA except for PGK1 (Supplementary Figure 2). This is because the expression intensity of PGK1 was too high and beyond the dynamic range of RNA-seq. Thus, the expression difference of PGK1 could not be calculated properly.

ROS generation in RCC4-EV and RCC4-VHL cells in response to propofol treatment. Next,
we investigated that impact of gene silencing of PDK1 on ROS generation and caspase 3/7 activity in response to propofol treatment in RCC4-EV cells. Lack of PDK1 gene expression induced ROS generation and caspase 3/7 activation (Fig. 7a,b). We found that ROS generation played a critical role in propofol-induced cell death. Here, we compared ROS levels in RCC4-VHL and RCC4-EV cells exposed to 50 µM propofol. ROS generation in response to 50 µM propofol treatment was observed in RCC4-VHL cells, but not in RCC4-EV cells (Fig. 7c). Moreover, treatment with the antioxidant NAC suppressed 50 µM propofol induced caspase 3/7 activation exclusively in RCC4-VHL cells (Fig. 7d,e). Next, we investigated the expression of genes which related to generation and scavenging ROS by RNA-Seq analysis. The expression levels of selected genes within GO:0016909 (antioxidant activity) (Fig. 7f) and GO:1903426-8 (regulation of reactive oxygen species biosynthetic process) (Fig. 7g) were investigated in these cell lines.
Effect of exogenous HIF-1 activation on propofol toxicity in neuronal SH-SY5Y cells. Finally, effect of exogenous HIF-1 activation on propofol-induced toxicity was examined in a different cell-type from RCC4 cells. Human neuroblastoma SH-SY5Y cells were treated with 100 µM nPG or 100 µM DMOG in the presence of 20% O 2 . The treatments induced expression of HIF-1α protein (Fig. 8a), suppressed OCR, and increased ECAR (Fig. 8b) in these cells. Consistent with our findings in RCC4-VHL cells, treatment with nPG or DMOG conferred resistance to the caspase 3/7 activation induced by 50 µM propofol (Fig. 8c) and cell death (Fig. 8d).

Discussion
In this study, we demonstrated that activation of HIF-1 by genetic or pharmacological means induced metabolic reprogramming and attenuated the ROS generation and cell death induced by a clinical relevant concentration  Figure 2. Propofol-induced caspases activation is attenuated in RCC4-EV cells than RCC4-VHL cells. RCC4-VHL and RCC4-EV cells were exposed to the indicated propofol concentrations for 6 h. The levels of (a) caspase 9 (n = 5) and (b) caspase 3/7 (n = 5) activity are shown for each treatment group. Differences between results were evaluated by two-way ANOVA followed by Dunnett's test for multiple comparisons; *p < 0.05, as compared to the control cell population (no treatment); # p < 0.05, for the indicated comparison. In RCC4 cells, the VHL gene is ablated 10 . Because VHL is an essential component of the E3 ubiquitin ligase, VHL regulates HIFα subunit protein expression 9,20 . HIF-1 is therefore activated in RCC4-EV cells under both 20% and 1% O 2 conditions. As demonstrated by the present RNA-Seq gene expression analysis, enrichment analysis, and qRT-PCR study, canonical glycolysis and the HIF-1-dependent pathway were activated in RCC4-EV cells. Thus the global gene expression analysis demonstrates change of metabolic mode from OXPHOS to glycolysis in RCC4-EV cells.
O 2 is primarily required for OXPHOS within cells. The maximal respiration rates in RCC4-EV cells were downregulated, as compared to the rates observed in RCC4-VHL cells. This indicated that mitochondrial electron transport was significantly inhibited in RCC4-EV, as compared to RCC4-VHL cells. Substrate availability is also a critical regulator of OXPHOS. While the most critical substrate for OXPHOS is O 2 , acetyl-CoA is another critical regulator of this process. The conversion of pyruvate to acetyl-CoA thus represents a critical regulatory point in cellular energy metabolism 17,21 . Pyruvate dehydrogenase is regulated by PDK-mediated phosphorylation of its E1 subunit. PDK1 is downstream of HIF-1 and it negatively regulates mitochondrial function by reducing pyruvate entry into the tricarboxylic acid cycle. The present study found that PDK1 mRNA expression increased in RCC4-EV cells to a greater extent than in RCC4-VHL or SH-SY5Y cells following treatment with HIFα-hydroxylase inhibitors. Suppression of PDK1 expression by siRNA increased ROS in response to propofol treatment. Proton leak, as defined by the mitochondrial respiration rate in the presence of oligomycin, was lower in RCC4-EV cells than in RCC4-VHL cells. Mitochondrial superoxide production is highly dependent on the membrane potential 22,23 and proton leak pathways may therefore minimize oxidative damage by reducing this potential and thus suppressing superoxide production. Together with metabolic reprogramming, HIF-1dependent gene expression change contributes to change of mode of electric transport in mitochondria.
Intriguingly, RCC4-EV cells had a higher intracellular ATP concentration than RCC4-VHL cells. This was consistent with the more active ECAR in RCC4-EV cells. ECAR provides a surrogate marker of glycolysis 21 , and the higher ECAR in RCC4-EV cells therefore indicated that this ATP was derived from glycolysis. We recently demonstrated that clinically relevant doses of propofol suppressed mitochondrial electron transport in SH-SY5Y  Data are expressed as the mean ± SD. Differences between results were evaluated by t-test; # p < 0.05 for the indicated comparison. (b) Caspase 3/7 activity in RCC4-EV and RCC4-VHL cells (n = 3), incubated with or without 50 µM propofol for 6 h, with or without 100 µM YC-1 as indicated. Differences between results were evaluated by two-way ANOVA followed by Dunnett's test for multiple comparisons; *p < 0.05, as compared to the control cells (no treatment); # p < 0.05 for the indicated comparison. (c) RCC4-VHL cells were incubated with or without 100 µM nPG or 100 µM DMOG for 24 h, as indicated, prior to determination of the indicated mRNA levels by qRT-PCR. Differences between results were evaluated by one-way ANOVA followed by Dunnett's test for multiple comparisons; *p < 0.05, as compared to the control cell population (no treatment). (d) Caspase 3/7 activity in RCC4-VHL cells (n = 3) that were exposed to the indicated treatments for 24 h prior to treatment with 50 µM propofol for 6 h. Differences between results were evaluated by two-way ANOVA followed by Dunnett's test for multiple comparisons; *p < 0.05, as compared to the control cell population (no treatment); # p < 0.05 for the indicated comparison.
SCIEnTIfIC RePORTS | (2018) 8:8987 | DOI:10.1038/s41598-018-27220-8 cells in a dose-and time-dependent manner 5 . Consistent with this conclusion, mitochondrial DNA-deficient cells were shown to be resistant to propofol-induced toxicity 5 . The reduction of O 2 to H 2 O by complex IV is not completely efficient. If electron transfer to O 2 occurs at complexes I or III, ROS generation occurs and these free radicals can oxidize cellular proteins, lipids, and nucleic acids. The ROS plays a critical role in propofol-induced cell death. In fact, the treatment with the antioxidant NAC reduced propofol-induced caspase 3/7 activation. Propofol induced significantly less ROS production in RCC4-EV cells than in RCC4-VHL cells. Propofol has also been shown to suppress the activity of complexes I, II and III, and to reduce mitochondrial oxygen consumption 5 . The RNA-Seq analysis indicated that HIF-1 activation on RCC4-EV cells induced gene expression which facilitate glycolysis but not significantly induced gene sets of peroxidase activity or reactive oxygen species metabolic process. Thus, HIF target gene activation is upstream of mitochondrial function and can alter mitochondrial activity.
Preclinical studies in animal models have predicted that systemic HIF activation has the potential to alter glucose, fat, and mitochondrial metabolism 24 . Indeed, a series of HIFα-hydroxylase inhibitors are currently undergoing evaluation in clinical anemia trials 25,26 . Thus, in addition to nPG and DMOG, these HIFα-hydroxylase inhibitors could modulate mitochondrial metabolism and may prevent the cell death that occurs during propofol infusion syndrome. The findings of the present study warrant a preclinical trial of these inhibitors for the treatment of this syndrome in an animal model.
In conclusion, VHL deletion or exposure to small-molecule HIFα-hydroxylase inhibitors activates HIF-1 and cellular metabolic reprogramming and oxygen utilization of mitochondria. The HIF-1 activation suppresses ROS generation and confers resistance to propofol toxicity.

Materials and Methods
Cell culture and reagents. Renal cell carcinoma cell lines were stably transfected with pcDNA3-VHL (RCC4-VHL) or empty vector (RCC4-EV) 6 . These cells lines were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 0.1 mg/ml streptomycin. The human neuroblastoma SH-SY5Y cells were maintained in RPMI medium supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 0.1 mg/ml streptomycin 6,27,28 . Purified mouse anti-human HIF-1α antibody Clone 54/HIF-1α was purchased from BD Biosciences (San Jose, CA). HIF-1β/ARNT (D28F3) XP rabbit monoclonal antibody was from Cell Signaling Technology (Danvers, MA). Dimethyloxaloylglycine (DMOG) and n-propyl gallate (nPG) the anti-β-actin antibody were obtained from Sigma. A list of reagents used in this study is provided in Table 1. AQueous One Solution ™ Reagent was added to each well and the plates were incubated at 37 °C for 1 h prior to measuring the absorbance of each sample using an iMark ™ Microplate Reader (BIO-RAD, Hercules, CA, USA) at a wavelength of 490 nm. Cell viability was then calculated by comparing the absorbance of the treated cells with that of the control cells (RCC4-VHL cells at 24 h incubation), which was defined as 100%. All samples were assayed in triplicate or quadruplicate for each experiment.

Reagents
Identifier Source Caspase activity assays. The levels of caspase 9 and caspase 3/7 activity were assessed using a Caspase-Glo ™ 9 Assay Kit (Promega) and an Apo-ONE ™ Homogeneous Caspase-3/7 Assay Kit (Promega), respectively, according to the manufacturer's protocols 6,28 . Briefly, cells were seeded into 96-well plates (2 × 10 4 cells/well) and incubated overnight. The following day, cells were treated with the indicated concentrations of the appropriate drug(s) for varying lengths of time. After treatment, 100 μl of Apo-ONE Caspase-3/7 Reagent ™ was added to each well. Cells were incubated at room temperature for 1 h and the luminescence of each well was measured using an EnSpire ™ Multimode Plate Reader (PerkinElmer, Waltham, MA, USA). Caspase activity was then calculated by comparing the levels of luminescence of the treated cells with that of the control cells (incubated without drugs), which was defined as 100%. Assays were performed in triplicate at least twice. Data were expressed as means ± the standard deviation (SD).
Immunoblot assays. Whole-cell lysates were prepared as described previously 6,27,29 . In brief, these were prepared using ice-cold lysis buffer containing 0.1% sodium dodecyl sulfate, 1% Nonidet P-40, 5 mM ethylenediaminetetraacetic acid, 150 mM NaCl, 50 mM Tris-Cl (pH 8.0), 2 mM dithiothreitol, 1 mM sodium orthovanadate, and Complete Protease Inhibitor ™ (Roche Diagnostics, Tokyo, Japan). Samples were centrifuged at 10,000 × g to sediment the cell debris, and the supernatant was used for subsequent immunoblotting experiments. For HIF-1α and HIF-1β determinations, 35 µg of protein was fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (7.5% gel), transferred to membranes, and immunoblotted using the indicated primary antibodies at a dilution of 1:500. Horseradish peroxidase-conjugated sheep anti-mouse IgG (GE Healthcare, Piscataway, NJ) was used as the secondary antibody, at a dilution of 1:2,000. The signal was developed using enhanced chemiluminescence reagent (GE Healthcare). Experiments were repeated at least three times and representative blots are shown.  with CellQuest Pro ™ software. Data were evaluated using FlowJo ™ version 7.6.3 software (TreeStar, Ashland, OR, USA), exported to Excel spreadsheets, and subsequently analyzed using the statistical application, Prism7 ™ . ATP assay. The CellTiter-Glo ™ luminescent cell viability assay kit (Promega, Madison, WI) was used to evaluate the intracellular ATP content 6 . Briefly, cells were seeded in 96-well plates (3 × 10 3 cells/well) and allowed to grow for 24, 48 and 72 h. CellTiter-Glo reagent (50 μl) was then added directly into each well and incubated for 10 min prior to reading the plate using an EnSpire ™ Multimode Plate Reader (PerkinElmer, Waltham, MA, USA).
This detected the luminescence generated by the luciferase-catalyzed reaction between luciferin and ATP. Assays were performed in triplicate at least twice. The ATP content was then calculated by comparing the luminescence levels of RCC4-VHL cells with that of RCC4-EV cells, which was defined as 100%. Data were expressed as the mean ± SD. The minimum OCR measured after rotenone/antimycin A injection was considered to represent the non-mitochondrial respiration rate. The basal OCR was calculated by subtracting the non-mitochondrial respiration rate from the last OCR measurement before oligomycin injection. The maximal OCR was calculated by subtracting the non-mitochondrial respiration rate from the maximum OCR measurement after FCCP injection. The proton leakage was calculated by subtracting the non-mitochondrial respiration rate from the minimum OCR measured after oligomycin injection.

Measurement of cellular oxygen consumption and extracellular acidification.
For XF Glycolysis Stress Test ™ for the ECAR, injection port A on the sensor cartridge was loaded with glucose (final concentration 10 mM), 2-Deoxy-D-glucose (final concentration 50 mM) loaded to portB and olygomycin (final concentration 1 µM) was loaded to port C. During the sensor calibration, cells were incubated at 37 °C in 180 μl assay medium (XF Base Medium and 2 mM l-glutamine, pH 7.4) in the non-CO 2 incubator. The plate was immediately placed into the calibrated XFp Extracellular Flux Analyzer for the assay.

Measurement of oxygen consumption in permeabilized cells.
The activity of individual respiratory chain complexes was evaluated in permeabilized cells 30,31 . Briefly, cells were washed with mitochondrial assay solution (MAS) buffer (220 mM mannitol, 70 mM sucrose, 10 mM KH 2 PO 4 , 5 mM MgCl 2 , 2 mM HEPES, 1 mM EGTA, 0.2% fatty acid-free bovine albumin, adjusted to pH 7.2 with KOH), and the medium was replaced with MAS buffer supplemented with 10 mM pyruvate, 1 mM malate, 4 mM ADP, and 1 nM plasma membrane permeabilizer ™ . The cells were then loaded into the XFp analyzer to measure respiration rates using cycles of 30 s mixing/30 s waiting/2 min measurement. Protocol A: After the measurement of pyruvate-driven respiration, rotenone (final concentration 2 µM) was injected through port A to halt the complex I-mediated respiratory activity. Next, succinate (10 mM) was injected through port B to donate electrons at complex II, bypassing complex I inhibition. The addition of antimycin A (2 µM) via port C inhibited complex III, and N,N, N′,N′-tetramethyl-p-phenylenediamine (TMPD 0.1 mM), combined with ascorbate (10 mM), was subsequently injected through port D to measure complex IV activity. Protocol B: As an alternative approach, cells were initially supplemented with pyruvate to measure complex I activity. After injection of rotenone, duroquinol was injected to stimulate complex III-mediated respiration.

Mitochondrial mass assay. Mitochondrial mass was measured by staining cells with MitoTracker ™ Green
FM at 37 °C for 15 min in PBS containing 5% FBS. Stained cells were filtered and analyzed immediately in a FACScan flow cytometer (BD Bioscience). Mean fluorescence intensity was analyzed using CellQuest software (BD Bioscience).

Semi-quantitative real-time reverse transcriptase-polymerase chain reaction analysis
(qRT-PCR). Total RNA was extracted from cells using the RNeasy ™ Mini Kit (Qiagen, Hilden, Germany), according to the manufacturer's instructions 6 . First-strand synthesis and RT-PCR were performed using the QuantiTect ™ Reverse Transcription Kit (Qiagen) and Rotor-Gene ™ SYBR Green PCR Kit (Qiagen), according to the manufacturer's protocol. Amplification and detection were performed using Rotor-Gene ™ Q (Qiagen). PCR primers were purchased from Qiagen. The change in expression of each target mRNA was calculated relative to the level of 18S rRNA.