In renal cell carcinoma, cancerous phenotypes linked to hypoxia-inducible factors are insensitive to the volatile anesthetic isoflurane

The possibility that anesthesia during cancer surgery may affect cancer recurrence, metastasis, and patient prognosis has become one of the most important topic of interest in cancer treatment. For example, the volatile anesthetic isoflurane was reported in several studies to induce hypoxia-inducible factors, and thereby enhance malignant phenotypes in vitro. Indeed, these transcription factors are considered critical regulators of cancer-related hallmarks, including “sustained proliferative signaling, evasion of growth suppressors, resistance to cell death, replicative immortality, angiogenesis, invasion, and metastasis.” We have now investigated the impact of isoflurane on the growth and migration of derivatives of the renal cell line RCC4. We confirmed that hypoxia-inducible factors do promote tumor growth and migration, but found that isoflurane does not affect cancerous phenotypes due to such factors.


Introduction
The hypothesis that anesthesia during cancer surgery may affect tumor recurrence, metastasis, and patient prognosis 1,2 is gaining increasing importance at present 3 .
Accordingly, there is a growing body of in vitro, in vivo, retrospective, and translational studies on the effect of anesthetics on perioperative immunity and cancer metastatic potential. For example, isoflurane was reported in several studies to induce hypoxia-inducible transcription factors (HIFs), and thereby enhance malignant phenotypes in vitro [4][5][6] . HIF-1 was originally cloned as a driver of erythropoietin expression 7-10 , but was linked shortly thereafter to tumor grade in various cancers 11 .
Indeed, HIFs are now well-known as critical regulators of cancer hallmarks, including "sustained proliferative signaling, evasion of growth suppressors, resistance to cell death, replicative immortality, angiogenesis, invasion, and metastasis" 12,13 . Moreover, tumor suppressors such as TP53 and PTEN also regulate HIFs. Another striking example of the physiological significance of HIFs is von Hippel-Lindau (VHL) disease, a hereditary cancer syndrome that predisposes to highly angiogenic tumors, in which the constitutive overexpression of vascular endothelial growth factor and glucose transporter 1 can be corrected by functional VHL protein, a tumor suppressor that targets HIFs for degradation. In this study, we investigated the impact of the volatile anesthetic isoflurane on growth and migration of derivatives of the renal cell line RCC4 that express (RCC-VHL) or do not express (RCC4-EV) VHL 14 . We demonstrate that HIFs significantly influence growth and migration, but isoflurane does not affect HIF-dependent phenotypes. -4-
As quantified by semi-quantitative RT-PCR, SLC2A1 (glucose transporter 1) and VEGFA (vascular endothelial growth factor A) were more abundant in RCC4-EV cells than in RCC4-VHL cells, but were induced in the latter at 1 % O 2 ( Figs. 2a and b).
However, expression in RCC4-VHL cells at 1 % O 2 was suppressed by isoflurane.
Effect of isoflurane on cell proliferation -5-Cell proliferation, as assessed by MTS assay, was higher in RCC4-EV cells than in RCC4-VHL cells, but was insensitive to isoflurane (Fig. 3a). Similarly, cellular ATP was more abundant in the former than in the latter, and was also insensitive to isoflurane (Fig. 3b).

Effect of isoflurane on cell migration
RCC4-EV cells migrated significantly faster than RCC4-VHL cells over 12 h (Fig. 4a), although exposure to isoflurane for 2 h significantly suppressed migration in both cells ( Fig. 4b).

Effect of isoflurane on glucose metabolism
In comparison to normal cells, cancer cells exhibit the Warburg effect, and thus preferentially metabolize glucose by glycolysis, producing lactate as an end product, despite availability of oxygen. Using an Extracellular Flux Analyzer™, the mitochondrial oxygen consumption rate was found to be lower in RCC4-EV cells in comparison to RCC4-VHL cells (Fig. 5a), but was insensitive to isoflurane (Fig. 5b).
On the other hand, extracellular acidification rate was higher in RCC4-EV cells relative to RCC4-VHL cells (Fig. 5c), but was also insensitive to isoflurane (Fig. 5d). Key parameters that determine the mitochondrial oxygen consumption rate, including basal oxygen consumption rate, maximum respiration, proton leak, and nonmitochondrial respiration, were also calculated from Cell Mito Stress Test™ data (Fig. 6). These parameters were significantly different between RCC4-EV and RCC4-VHL cells, but were insensitive to isoflurane.

Effect of isoflurane on global gene expression
Clustering of RNA sequencing data (Supplementary Information Table1) indicated that transcriptomic bias due to isoflurane was smaller than transcriptomic variations due to VHL expression (Fig. 7a). Indeed, more than 200 genes were differentially expressed between RCC4-EV and RCC4-VHL cells, as inferred from Wilcoxon signed rank test of FPKM values at significance level 0.05 (Fig. 7b). However, only one gene was differentially expressed in RCC4-VHL cells exposed to isoflurane, while no such gene was identified in RCC4-EV cells. Pairwise scatter plots comparing log 10  Finally, only 42 genes annoted to cancer hallmark gene ontologies were sensitive to isoflurane, although the effects were negligible (Fig. 8). Indeed, only CITED1 was strongly responsive to isoflurane.

Discussion
Collectively, the data suggest that clinically relevant doses of isoflurane suppress cell mobility, HIF-dependent intracellular signaling, and expression of genes associated with cancer hallmarks and phenotypes. We also confirmed that HIFs sustain cancer-associated gene expression, metabolism, cell proliferation, and cell motility.
The hallmarks of cancer were originally proposed by Hanahan and Weinberg in 2000 12 , and included sustained proliferative signaling, evasion of growth suppressors, resistance to cell death, replicative immortality, angiogenesis, invasion, and metastasis.
Subsequent conceptual progress has added another two emerging hallmarks, namely reprogramming of energy metabolism and evasion of immune destruction 13 .
Accordingly, we investigated the impact of isoflurane and HIF on these hallmarks based on global gene expression. Noting that HIF has been extensively investigated in the context of cancer biology 4,11 , we used RCC4-EV cells, which are derived from human renal cell carcinoma. As these cells are VHL-deficient, both HIF-1 and HIF-2 are activated even under normoxic conditions, but are suppressed by forced expression of VHL, as in RCC4-VHL cells 14 . Accordingly, RCC4-EV cells proliferate and migrate faster than RCC4-VHL cells, and exhibit metabolic reprogramming from oxidative phosphorylation to glycolysis. These results clearly indicate that HIFs are critically involved in cancerous phenotypes, as previously reported.
However, isoflurane treatment did not affect HIF-1α and HIF-2α expression in RCC4-EV cells. In contrast, isoflurane suppressed HIF-1α expression in RCC4-VHL cells at 1 % O2. Accordingly, isoflurane also suppressed SLC2A1 and VEGFA, which are downstream of HIF-1, under normoxic or hypoxic conditions, although this effect is far smaller than that of VHL expression. Further, isoflurane also slowed growth and -8-migration, but also to a smaller extent than VHL expression. Strikingly, exposure to isoflurane for 2 h was also demonstrated to boost HIF-1α and HIF-2α protein expression within 6 h in RCC4 cells, and in a PTEN/Akt-dependent manner 15 . Indeed, HIF-1α and HIF-2α were barely detectable in that study without isoflurane treatment, contradicting for unknown reasons several other reports, including the first to link isoflurane to HIFs 14,[16][17][18] . Hence, we investigated potential protocol-dependent effects, but found that neither exposure to isoflurane for 8 h, nor for 2 h followed by culture for another 6 h, induced HIF-1 and HIF-2 in RCC4-EV cells.
In 2006, Exadaktylos et al. 2 proposed that anesthesia and analgesia during cancer surgery may affect tumor recurrence or metastasis, a hypothesis that was subsequently supported by several clinical studies 1 . Potential underlying mechanisms include direct cellular effects, as well as indirect effects on patient immunity and on cancer metastasis.
On the contrary, we found that isoflurane clearly has adverse effects against cancer cells, isoflurane suppressing migration but not proliferation, both of which are linked to HIFs, and which may determine the fate of cancer cells. In addition, metabolism and global gene expression appeared to be sensitive to HIFs but not to isoflurane. We note, however, that our data are entirely from in vitro experiments in established cell lines, while xenografts may be required to elucidate the impact of anesthetics on cancer progression in vivo.
In summary, we demonstrated that isoflurane does not affect HIF activity in renal carcinoma cells, nor the expression of genes associated with cancer hallmarks.

Hypoxic and isoflurane treatment
Cells were maintained in an airtight chamber or workstation (AS-600P; AsOne, Osaka,  Rotor-Gene™ SYBR Green PCR Kit (Qiagen), following the manufacturer's protocol. non-mitochondrial oxygen consumption rate = minimum rate after rotenone/antimycin A injection; proton leak = minimum rate after oligomycin injection -non-mitochondrial respiration. To measure extracellular acidification rate, injection port A was loaded with -13-10 mM glucose, and the sensor was calibrated with cells incubated at 37 °C in a non-CO 2 incubator and in 180 μ L of assay medium (XF base medium with 2 mM L-glutamine, pH 7.4). The plate was immediately assayed following calibration and loading with oligomycin (1 µM) and 50 mM 2-deoxy-D-glucose. Extracellular acidification rate was normalized to total protein/well and calculated as extracellular acidification rate (glycolysis) = maximum rate after glucose injection − last rate before glucose injection.

Cell migration assay
Cell migration was analyzed using Oris™ Cell Migration Assay (Platypus Thechnologies, Madison, WI). Cells (2 × 10 4 ) were seeded in wells plugged with stoppers to restrict seeding to outer areas only. Cells were then exposed for 8 h to 21 % oxygen and 5 % carbon dioxide balanced with nitrogen with or without 2 % isoflurane.
Stoppers were then removed to expose unseeded sites, into which cells could migrate during subsequent incubation at 37 °C in 5 % CO 2 and 95 % air for indicated times. Cell migration was imaged on a BZ-9000 Fluorescence Microscope (KEYENCE, Itasca, IL), and colonized areas were quantified in pixels in ImageJ 1.51 (National Institutes of Health), corrected for total unseeded area, and expressed as percentage of colonized areas in reference wells.

RNA sequencing
Total RNA was extracted from cells using RNeasy Mini Kit (Qiagen) 19,25 , and FASTQ files were evaluated as described previously 19,25 . In brief, gene lists for Metascape analysis were generated from the Cuffdiff output file of significantly differentially expressed genes (p < 0.05; Table S1). Gene ontology annotations were