Midazolam and Dexmedetomidine Affect Neuroglioma and Lung Carcinoma Cell Biology In Vitro and In Vivo

SURGICAL removal of primary tumors is often the first-line treatment for most cancers. Unfortunately, potentially curative surgical resection may paradoxically open a window of opportunity whereby residual cancer cells can overcome host defenses and become established as distant metastatic disease that can prove fatal.¹  The perioperative period has been established as a critical phase in influencing long-term cancer outcomes, with drugs used for general anesthesia implicated in affecting the behavior of tumor cells.^2–4 

Dexmedetomidine is a widely used sedative agent, with analgesic, sympatholytic, and anxiolytic effects during the perioperative period. It is a highly selective and potent α₂-adrenoceptor agonist with dose-dependent effects.⁵  Dexmedetomidine has been shown to induce antiapoptotic effects through α₂-adrenoceptor signaling in neurons and immune cells, which are concerning for the use of the drug in the context of cancer.^6–9  Recent studies on the postoperative effect of dexmedetomidine on metastatic progression in vivo suggest that it is associated with reduced overall survival rates, although the molecular mechanisms that mediate this are not yet fully understood.^10,11 

The expression of α₂-adrenoceptors in breast cancer have been associated with increased tumor growth and migration both in vitro and in vivo, with activation of extracellular signal-related protein kinases by α₂-adrenoceptor agonists such as dexmedetomidine and clonidine shown to enhance this effect in vitro in breast cancer cells.^12–15  Inactivating α₂-adrenoceptor signaling inhibited the proliferation, migration, and invasion of breast cancers.¹⁶  In lung alveolar epithelial cells, dexmedetomidine was found to protect against the effects of apoptosis induced by oxidative stress in vitro by enhancing cell survival and proliferation.¹⁷  However, its functional and mechanistic effects have yet to be investigated in other types of cancers.

Midazolam is a widely used anesthetic from the benzodiazepine class, with a wide variety of uses including premedication, procedural sedation, and the treatment of insomnia and seizures. It exerts its action through benzodiazepine binding sites of the γ-aminobutyric acid type A (GABA_A) receptor, which mediates its primary clinical effects. Midazolam also binds peripheral benzodiazepine receptors, which tend to be localized to the outer mitochondrial membrane.^18,19 

Several types of cancers, including breast, ovarian, liver, and colon, show increased peripheral benzodiazepine receptor expression relative to healthy tissue.^20–22  In the breast cancer cell lines examined, expression of this receptor was relatively higher in the more aggressive phenotypes. Peripheral benzodiazepine receptors regulate various cellular functions including proliferation, oxidative processes, and apoptosis, highlighting their role in mitochondrial function.¹⁸  Midazolam has been reported to activate the apoptotic extrinsic death receptor or intrinsic mitochondrial pathways to promote cellular apoptosis, but further underlying molecular mechanisms remain elusive.^23,24 

In the current work, we hypothesized that dexmedetomidine enhances cell proliferation, survival, and migration of human neuroglioma (H4) and lung carcinoma (A549) cell lines, whereas midazolam has apoptosis-inducing activity and suppresses cancer cell progression. We investigated changes in cell cycle and apoptotic markers, mitochondrial function, and cell signaling by α₂-adrenoceptors and benzodiazepine receptors, caused by dexmedetomidine and midazolam, respectively. The potential effect of dexmedetomidine and midazolam on tumor burden were also investigated in xenograft mice.

Human lung carcinoma cells (A549) were obtained from the European Collection of Authenticated Cell Cultures (United Kingdom) and were authenticated by short-tandem repeat polymerase chain reaction DNA profiling. Human neuroglioma cells (H4) were a kind gift from Zhongcong Xie, M.D., Ph.D. (Massachusetts General Hospital, Boston, Massachusetts). H4 cells were purchased originally from and authenticated by the American Tissue Culture Collection by short-tandem repeat analysis. A549 and H4 cells were cultured at 37°C in a humidified atmosphere of 5% CO₂. A549 cells were maintained in Roswell Park Memorial Institute-1640 medium and H4 cells in Dulbecco’s Modified Eagle Medium (Sigma-Aldrich, United Kingdom), supplemented with 10% fetal calf serum (HyClone, New Zealand), 1% l-glutamine, and 1% penicillin-streptomycin (Sigma-Aldrich). Cell morphology and behavior were closely monitored throughout the duration of study to ensure no phenotypic changes and contamination.

Dexmedetomidine hydrochloride (Sigma-Aldrich) was used at concentrations of 0.001, 0.01, 0.1, 1, and 10 nM.¹⁴  Midazolam (Sigma-Aldrich) was used at concentrations of 0.01, 0.1, 1, 10, 25, 50, 100, 200, and 400 μM.²⁵  Isotonic sodium chloride solution was used as the vehicle control for both anesthetics. The α₂-adrenoceptor antagonist atipamezole hydrochloride (Antisedan; Elanco Animal Health, United Kingdom) was used at a concentration of 10 nM, and the benzodiazepine receptor antagonist flumazenil (Sigma-Aldrich) was used at a concentration of 40 mM; antagonists were used at 10× greater concentrations, identified from our pilot studies, to ensure complete saturation of relevant receptors. Cells were exposed to antagonists for 10 min and then washed before addition of anesthetic agents in all combined experiments.

Cell viability was assessed with colorimetric 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyl-2H-tetrazolium bromide assay (EMD Chemicals, USA). Culture medium was removed after incubation with or without indicated concentrations of drug. Cells in 24-well plates were treated with 500 µl 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyl-2H-tetrazolium bromide (0.5 mg/ml) and then incubated for 4 h under normal culture conditions. 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyl-2H-tetrazolium bromide was carefully removed; 500 μl of dimethyl sulfoxide (Fisher Scientific, United Kingdom) was added per well and then left for 10 min to allow dissolution of the crystals into a homogenous purple color. Forty microliters of this solution was transferred into a 96-well plate (VWR, United Kingdom), and a further 160 µl dimethyl sulfoxide was added. Absorbance was analyzed with a spectrophotometer (MRX Microplate Reader; Dynex Technologies, USA) at a wavelength of 595 nm. For additional experiments with midazolam at lower concentrations, cell viability was assessed using the Cell Counting Kit-8 assay. A549 and H4 cells were seeded into 96-well plates. Once confluent, cells were treated with 0.01, 0.1, 1, or 10 μM midazolam. Into each well, 10 μL of Cell Counting Kit-8 solution (Sigma-Aldrich) was added, and then the cells were incubated for 1 h under normal culture conditions. Absorbance was read at 450 nm with an ELx800 Microplate Reader (BioTek, USA). Cell viability was expressed relative to the control.

Cell migration was assessed with the wound-healing assay in vitro. Cells were cultured in 60-mm Petri dishes overnight to form a confluent monolayer. The scratch was performed on this monolayer with a 1-ml pipette tip under sterile conditions, after which interventions were added. After treatment, cells were washed twice with culture media. Images were taken with a digital camera (Canon, United Kingdom) immediately after addition of drugs, at 24 h, and at 48 h. Marks were made at the bottom of the dish to ensure that all images were taken at the same site. The area within the scratch occupied by cells at the different time points was calculated with Image-Pro Plus software (Media Cybernetics, USA). The percentage of gap closure was used as an indicator of cell migration.

A549 or H4 cells were fixed in 4% paraformaldehyde immediately after treatment and then blocked with 10% normal donkey serum (Sigma-Aldrich) for 1 h, followed by overnight incubation at 4°C in primary antibodies: mouse anti-Ki-67 (Dako, United Kingdom); anticyclin A; anticyclin D; anticyclin E; antiBcl-2; antiBcl-xL; antiBad; antimammalian target of rapamycin (Abcam, United Kingdom); rabbit anticleaved caspase 3, 8, or 9; anti-α₂-adrenoceptor; or antibenzodiazepine receptor (Abcam). Cells were washed with phosphate-buffered saline and then incubated in fluorescein isothiocyanate or rhodamine-conjugated secondary antibody (Merck Millipore, United Kingdom). Cells were counterstained and mounted with nuclear dye 4′, 6-diamidino-2-phenylindole–mounting medium (Vector Laboratories, USA). Images were captured under an Olympus BX4 microscope (United Kingdom). Fluorescent intensity was quantified with the mean pixel intensity of relevant antibody staining by ImageJ software (National Institutes of Health, USA). Ten representative regions per section were randomly selected by an assessor blinded to the treatment groups. Intensity values were calculated and expressed as relative to control.

JC-1 is a membrane-permeable dye used in apoptotic studies that exhibits potential-dependent accumulation in mitochondria, indicated by the shift from green fluorescent monomers to red fluorescent J-aggregates. A549 or H4 cells were cultured in 24-well plates at 37°C with or without anesthetic treatment. To each well, 100 μl of JC-1 staining solution (Abcam) per milliliter of culture medium was added. The solution was then gently agitated before incubation at 37°C in the dark for 15 min. Red fluorescent J-aggregates and green fluorescent JC-1 monomers were detected at excitation/emission wavelengths of 540/570 nm and 485/535 nm, respectively. Mitochondrial depolarization is indicated by a decrease in the red/green fluorescence intensity ratio.

Intracellular reactive oxygen species generation was detected by flow cytometry with the fluorescent dye dichlorodihydrofluorescein diacetate (Sigma-Aldrich). Cells were briefly washed twice with prewarmed Dulbecco’s phosphate-buffered saline (Sigma-Aldrich). To stain cells, 10 µM dichlorodihydrofluorescein diacetate was added into a fluorescence-activated cell sorting tube in the dark for 30 min at 37°C. Immunofluorescence intensity was acquired and analyzed with flow cytometry (FACSCalibur; Becton Dickinson, USA). Each trial had at least 10,000 gated events.

Adult male BALB/C immunodeficient nude mice (Harlan, United Kingdom) were bred in temperature- and humidity-controlled cages in a specific pathogen-free facility at Chelsea and Westminster Campus, Imperial College London (United Kingdom). This study was approved by the Home Office, United Kingdom; all animal procedures were conducted in accordance with the United Kingdom Animals (Scientific Procedures) Act of 1986. A549 tumor cells (approximately 3.75 × 10⁶) were suspended in 100 μl Roswell Park Memorial Institute-1640 media and injected subcutaneously into the middle left side of the mice. Four days after tumor cell implantation, mice were randomly selected and placed into one of three separate cages for treatment. Mice were treated daily for five days with saline, clinically relevant doses of midazolam (2.5 mg/kg per day), or dexmedetomidine (0.5 mg/kg per day) via subcutaneous injection in a different site to tumor implantation (n = 6 per group).²⁶  To ensure no cardiorespiratory disturbances by drug administration, each single dose of both sedatives given did not induce a sedative state to the animals. No statistical power calculation was conducted before the study given that the sample size was based on availability for this proof-of-concept study, and all data obtained have been reported. Three weeks after tumor implantation and treatment, mice were euthanized and underwent ascending aortic perfusion with 4% paraformaldehyde. Subdermal solid tumor masses were excised from mice and stored in 30% sucrose solution at 4°C to allow dehydration. Tumor tissues were embedded in optimal cutting temperature solution (Tissue-Tek, Japan) and cryosectioned into 15µm thick sections. Frozen sections were blocked in 10% normal donkey serum at room temperature for 1 h and then incubated overnight at 4°C in primary antibodies: rabbit monoclonal cyclin-D1 antibody (Abcam) or rabbit polyclonal Ki67 antibody (Santa Cruz Biotechnology, USA). Slides were then incubated the following day in donkey antirabbit FITC (Merck Millipore) secondary antibody for 2 h at room temperature. Slides were counterstained and mounted with 4′, 6-diamidino-2-phenylindole–mounting medium and viewed under Olympus BX4 microscope.

Data were first tested for normality with the Shapiro–Wilk test and then analyzed by one- or two-way ANOVA and Tukey post hoc test with GraphPad Prism (Version 7.0; GraphPad Software, USA). Only the following data were analyzed with two-way ANOVA: all cell viability assays and the wound-healing assay performed with dexmedetomidine in combination with inhibitor atipamezole. One-way ANOVA analysis factor was “treatment” between subjects. Two-way ANOVA analysis factors for the cell viability assays were “cell line” within subjects and “treatment” between subjects; for the wound-healing assay, factors were “time” within subjects and “treatment” between subjects. Data are presented as bar charts or scatter plots and expressed as mean ± SD. No outliers were detected and no data were excluded from the analysis. Differences with P value less than 0.05 were considered to be statistically significant.

To examine whether dexmedetomidine and midazolam affect the growth of human lung carcinoma A549 cells and neuroglioma H4 cells, the 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyl-2H-tetrazolium bromide assay was used to measure cell viability. It was found that dexmedetomidine treatment for 24 h increased cell viability from concentrations as low as 0.001 nM (1.2-fold for both A549 [P = 0.007] and H4 cells [P < 0.0001] vs. vehicle) up to 10 nM (1.7-fold for A549 [P < 0.0001] and 1.6-fold for H4 cells [P < 0.0001] vs. vehicle; fig. 1A). However, midazolam treatment for 24 h decreased cell viability in a dose-dependent manner, from the highest concentration of 400 μM (2.6-fold for A549 and 1.9-fold for H4 cells vs. vehicle, P < 0.0001) down to 25 μM (1.2-fold for both A549 and H4 cells vs. vehicle, P < 0.0001; fig. 1B). Midazolam treatment at lower concentrations between 0.01 μM and 10 μM had no effect on cell viability, assessed using the Cell Counting Kit-8 assay (Supplemental Digital Content, fig. S1, https://links.lww.com/ALN/B760). To further investigate whether the observed changes in cell viability were due to cell death or proliferation, Ki67 expression was assessed through immunofluorescence staining in A549 and H4 cells treated with 1 nM dexmedetomidine (fig. 1, C and D) or 400 μM midazolam (fig. 1, E and F). The Ki67 protein is strictly associated with cellular proliferation and is present during all active phases of the cell cycle. Dexmedetomidine increased Ki67 expression in both A549 and H4 cells (2.9- and 2-fold, respectively, vs. vehicle, P < 0.0001). However, midazolam decreased Ki67 expression (1.7-fold for A549 [P < 0.0001] and 1.4-fold for H4 cells [P = 0.0003] vs. vehicle).

The effects of dexmedetomidine treatment on migration of A549 and H4 cells, and of midazolam on cell migration of A549 cells, were assessed with the wound-healing assay (fig. 2). Our results indicate that migration of both cell lines is enhanced by dexmedetomidine (by 2.2-fold in A549 and 1.9-fold in H4 cells vs. vehicle, P < 0.0001) but decreased by midazolam (by 2.6-fold in A549 cells vs. vehicle, P < 0.0001).

To observe which aspects of the cell cycle are affected by dexmedetomidine and midazolam, A549 and H4 cells treated with the two agents were stained for cyclin A (fig. 3), D, and E (Supplemental Digital Content, fig. S2, https://links.lww.com/ALN/B761, and S3, https://links.lww.com/ALN/B762). Similar trends were observed for all three markers, with dexmedetomidine increasing expression compared with the vehicle control: cyclin A by at least 1.8-fold in both cell lines, cyclin D by 1.8- and 2.2-fold in A549 and H4 cells, respectively, and cyclin E by 2.6- and 2.1-fold in A549 and H4 cells, respectively. In contrast, midazolam decreased cyclin expression compared with the vehicle control: cyclin A by 1.4- and 1.5-fold in A549 and H4 cells, respectively, cyclin D by 1.5-fold in both cell lines, and cyclin E by 1.6-fold in both cell lines. All observations were statistically significant (P < 0.001).

As previously observed, dexmedetomidine and midazolam affected cell viability differently. The effects of both agents on cell death were first investigated through immunofluorescence staining of antiapoptotic proteins Bcl-2 and Bcl-xL in H4 cells (Supplemental Digital Content, fig. S4, https://links.lww.com/ALN/B763). Dexmedetomidine increased expression of Bcl-2 and Bcl-xL (2.9- and 2.4-fold, respectively, vs. vehicle, P < 0.0001), whereas midazolam decreased expression of both Bcl-2 (P = 0.0002) and Bcl-xL (P = 0.002) by 1.6-fold compared with the vehicle. To corroborate this and identify the cell death response involved, we evaluated concentrations of cleaved caspases 3, 8, and 9. No differences in any caspases were detected with dexmedetomidine treatment (Supplemental Digital Content, fig. S5, https://links.lww.com/ALN/B764, and S6, https://links.lww.com/ALN/B765). It is interesting that midazolam caused the activation of caspases 3 and 9 (fig. 4), but not caspase 8 (Supplemental Digital Content, fig. S6, https://links.lww.com/ALN/B765), suggesting stimulation of the intrinsic mitochondrial apoptotic pathway. Reactive oxygen species, cytochrome C, and Bad play an important role in this pathway. Intracellular concentrations of reactive oxygen species were measured in both A549 and H4 cells after midazolam treatment with flow cytometry, and it was found that midazolam treatment increased reactive oxygen species concentrations in A549 (P = 0.0003) and H4 (P = 0.0002) cell lines by at least 1.1-fold compared with the vehicle treatment (fig. 5, A and B). Cytochrome C and Bad were also found to increase after midazolam treatment in H4 cells (2.3- and 2.5-fold, respectively, vs. vehicle, P < 0.0001; Supplemental Digital Content, fig. S7, https://links.lww.com/ALN/B766). In addition, midazolam caused depolarization of the mitochondrial membrane, as reflected by the decreased amount of J-aggregates (red fluorescent staining) compared with JC-1 monomers (green fluorescence), indicated by a lower red/green ratio compared with the vehicle control in A549 (2.3-fold, P = 0.0004) and H4 cells (3.2-fold, P = 0.0005; fig. 5, C and D).

For the purpose of understanding the specificity of dexmedetomidine and midazolam and their actions via the α₂-adrenoceptor and benzodiazepine receptor, respectively, studies were carried out with atipamezole, a synthetic α₂-adrenoceptor antagonist, and flumazenil, a benzodiazepine receptor antagonist. The presence of these receptors was first confirmed in both A549 and H4 cell lines (Supplemental Digital Content, fig. S8, https://links.lww.com/ALN/B767, and fig. S9, https://links.lww.com/ALN/B768). It was found that atipamezole could reduce dexmedetomidine-enhanced cell migration of A549 cells at 24 h and 48 h to baseline concentration (Supplemental Digital Content, fig. S8B, https://links.lww.com/ALN/B767). Atipamezole was also able to decrease the cell viability-promoting effects of dexmedetomidine, although atipamezole alone increased cell viability of both A549 and H4 cells (Supplemental Digital Content, fig. S8C, https://links.lww.com/ALN/B767). Further to this, atipamezole alone had no effect on cyclin A or D expression. However, it partially reduced dexmedetomidine-enhanced cyclin A and D expressions (Supplemental Digital Content, fig. S8D and S8E, https://links.lww.com/ALN/B767). Dexmedetomidine also significantly increased mammalian target of rapamycin expression in A549 cells by 1.5-fold compared with the vehicle control (P = 0.0006; Supplemental Digital Content, fig. S7C, https://links.lww.com/ALN/B766). However, flumazenil alone had no effect on cell viability and Ki67 expression but partially reversed the decreases in cell viability and Ki67 expression caused by midazolam treatment in both A549 and H4 cells (Supplemental Digital Content, fig. S9, https://links.lww.com/ALN/B768).

To extrapolate our in vitro findings to an in vivo setting, we treated nude mice implanted with A549 tumors with saline, dexmedetomidine, or midazolam. No differences in animal weight were detected before euthanizing the mice (fig. 6A); however, midazolam significantly reduced tumor size compared with tumors from naïve animals (fig. 6, B and C). Tumor tissue sections stained for Ki67 (fig. 6D) and cyclin D (fig. 6E) showed a decrease in both protein expressions in midazolam-treated animals and an increase in both markers in dexmedetomidine-treated animals, compared with animals without treatment. We noted that three weeks after tumor implantation, two of six dexmedetomidine-treated mice developed skin ulcers on the xenograft and reached clinical endpoints evaluated with a well-established scoring system.

In the present study, dexmedetomidine enhanced the proliferation and migration of A549 and H4 cells in vitro via adrenergic signaling and upregulation of antiapoptotic proteins Bcl-2 and Bcl-xL. In contrast, midazolam suppressed cancer cell proliferation and migration, induced apoptosis via the intrinsic mitochondrial pathway, caused mitochondrial depolarization, and increased intracellular reactive oxygen species in vitro. Midazolam reduced tumor burden, Ki67 expression, and cyclin D expression in vivo. Although dexmedetomidine did not affect tumor growth in vivo in this particular scenario, Ki67 and cyclin D expression were both found to be upregulated in extracted tumor tissue.

The effects of adrenergic signaling in breast cancer have been reported, and our study on dexmedetomidine suggests that these effects may extend to other cancer types.^14,27,28  Midazolam has been identified to have apoptosis-inducing effects in several cancer cell lines in vitro, including MA-10 Leydig cells, lymphoma, neuroblastoma, and leukemia.^24,25,29  We found that dexmedetomidine increased, and midazolam decreased, A549 and H4 cell viability in a dose-dependent manner. We consistently observed an increase in Ki67 and cyclin A, D, and E expression after dexmedetomidine treatment. Ki67 is present during all phases of the cell cycle (G₁, S, G₂, and mitosis) but absent from resting cells (G_O), providing a reliable marker of growth.³⁰  Cyclins regulate cell cycle progression, each protein associated with a different stage of the cycle. A consistent up- or downregulation in Ki67 and all cyclins assessed in this study shows the broad rather than specific effects of the drug treatment on the cell cycle. Atipamezole, an α₂-adrenoceptor antagonist, reversed the effects of dexmedetomidine on cell proliferation and migration entirely but only partially reduced cyclin A and D induction. This suggests that dexmedetomidine may influence pathways other than that activated by α₂-adrenoceptors. Flumazenil partially reversed midazolam-induced decreases in cell viability and proliferation. Flumazenil is a specific competitive antagonist at benzodiazepine receptors associated with receptors for GABA_A, indicating that midazolam effects were partly mediated by the benzodiazepine binding site of GABA_A or peripheral benzodiazepine receptors.³¹  Indeed, our data clearly show that benzodiazepine receptors exist in both A549 and H4 cell lines. Other mechanisms, including intracellular acidosis induced by benzodiazepines, may also contribute to the “killing” effect of midazolam.³² 

The two main pathways of apoptosis for cells are extrinsic or intrinsic. The extrinsic pathway is initiated by the death receptor, activating the death-inducing signaling complex and then caspase 8. The intrinsic pathway requires permeabilization of the mitochondria and release of cytochrome C, initiating the caspase cascade by activating caspase 9. This pathway can be triggered by cytochrome C and reactive oxygen species.^33,34  Both pathways converge onto effector caspase 3.³⁵  Here, midazolam induced apoptosis in both A549 and H4 cells through the intrinsic pathway, demonstrated by upregulated cleaved caspase 9 and 3, but not cleaved caspase 8, in addition to increased concentrations of cytochrome C and reactive oxygen species.

In cancer cells, high concentrations of reactive oxygen species are produced due to various factors such as increased metabolic activity, mitochondrial dysfunction, and cellular signaling. Reactive oxygen species can enhance cellular proliferation and migration and induce DNA damage that potentiates tumorigenesis. However, they can also induce cellular senescence and cell death. Reactive oxygen species production by carcinoma cells tends to be elevated in vitro and in vivo.³⁶  Midazolam has previously been shown to induce apoptosis possibly through the suppression of reactive oxygen species production in K562 human leukemia cells.²⁴  Our study, however, showed midazolam increased reactive oxygen species in both A549 and H4 cells. Excessive concentrations of intracellular reactive oxygen species can induce cell cycle arrest, senescence, apoptosis, and even uncontrolled cell death.^34,37,38  Increased mitochondrial oxidative stress causes cytochrome C release, the activation of caspases, and cell death, all observed in this study.^34,39  Mitochondrial release of free radicals leads to activation of the c-Jun N-terminal kinase signaling pathway. Upon reactive oxygen species stimulation, c-Jun N-terminal kinases phosphorylate and cause the downregulation of antiapoptotic proteins including Bcl-2 and Bcl-xL, both of which have been shown to protect cells from apoptosis induced by reactive oxygen species.^39–41  Our data collectively suggest that midazolam stimulates the intrinsic mitochondrial pathway, through a disproportional increase in reactive oxygen species.

Although midazolam is more widely reported to act via the GABA_A receptor, our data show midazolam caused mitochondrial depolarization in both cell lines investigated, supporting the notion that midazolam also acts via peripheral benzodiazepine receptors on the outer mitochondrial membrane.^18,19  These receptors are part of the mitochondrial permeability transition pore, located at the contact site between the inner and outer mitochondrial membranes.⁴²  The opening of this pore is a critical event during apoptosis and can be triggered by pseudopathologic conditions of oxidative stress in vitro.⁴³  Partial permeabilization of the inner mitochondrial membrane leads to an abrupt loss in transmembrane potential, whereas complete permeabilization of the outer mitochondrial membrane results in leakage of potentially toxic intermembrane proteins that result in apoptosis.⁴⁴  Endozepine, the endogenous ligand of peripheral benzodiazepine receptors, released from damaged mitochondria may theoretically act on intact mitochondria, participating in a positive feedback loop to accelerate the induction of mitochondrial membrane permeabilization.^45,46  Thus, we suggest two potential mechanisms for the action of midazolam as observed in this study: midazolam may bind to the peripheral benzodiazepine receptors located on the mitochondria to stimulate cellular apoptosis and/or may trigger the opening of the mitochondrial permeability transition pore to induce apoptosis by increasing intracellular reactive oxygen species. These hypotheses certainly warrant further studies.

Dexmedetomidine increased cancer cell survival by modulating Bcl-2 and Bcl-xL expression. Cancer cells often upregulate antiapoptotic Bcl-2 proteins as a protective mechanism against apoptosis, preventing mitochondrial outer membrane permeabilization.³³  Overexpression of Bcl-2 alone is not oncogenic, but it can dramatically enhance tumorigenesis in combination with other growth-promoting oncogenes such as Myc.⁴⁷  Further to this, we observed an increase in mammalian target of rapamycin expression, showing the potential of dexmedetomidine to perturb cellular signaling. The mechanism of α₂-adrenergic receptors includes the activation of inhibitory G-proteins or modulation of ion channel activity, although its effects on specific signaling pathways are still unclear.⁴⁸  The phosphoinositide-3-kinase/Akt/mammalian target of rapamycin pathway is critical in the regulation of numerous cellular functions including proliferation, survival, and differentiation and is one of the most activated pathways in cancers.⁴⁹  Mammalian target of rapamycin is a serine/threonine kinase that monitors nutrient availability, cellular oxygen and energy levels, and mitogenic signals.⁵⁰  Dexmedetomidine increases phosphorylated Akt and mammalian target of rapamycin concentrations as a mechanism of inhibiting neuronic autophagy.⁵¹  In a previous study, we found that dexmedetomidine could reverse the downregulation of phosphorylated mammalian target of rapamycin in lung alveolar epithelial cell injury induced by hydrogen peroxide.¹⁷  In this study, we observed prolonged dexmedetomidine treatment alone could increase mammalian target of rapamycin expression in the same cell line. Thus, dexmedetomidine can promote tumorigenesis through more than one mechanism including enhancing cell cycle progression and cell survival; however, further work is required on its effects on cancer cell signaling pathways.

In line with our in vitro data, our in vivo work shows that midazolam treatment reduces tumor size, Ki67 expression, and cyclin D expression in mice. Dexmedetomidine increased the expression of both these proteins, but surprisingly, we did not observe any significant effects on tumor size. A recent in vivo study found that dexmedetomidine administered by a slow release vehicle, maintaining prolonged exposure to the drug, increased tumor cell retention and growth of metastasis of breast, lung, and colon cancer cells in rodent models.¹¹  It is likely that the living duration (three weeks) given to mice after tumor implantation in our study was not sufficient to allow any significant changes in tumor growth caused by the treatment regime given. The reason this duration was set up is because the tumor burden in mice treated with dexmedetomidine caused severe whole body physiologic deterioration in at least two animals and with animal welfare in mind, all animals were euthanized at the end of the third week after tumor implantation, limiting our in vivo study.

Another limitation of this study appears to be the high doses used for treatment. Although the midazolam and dexmedetomidine doses used in our experiments have been reported in previous in vitro studies, caution needs to be taken when interpreting this data, as experimental in vitro and in vivo conditions cannot accurately represent clinical conditions.^14,24,25,52  The clinical concentrations for dexmedetomidine range from 0.2 to 2 ng/ml (1 to 10 nM), and the concentrations used in this study correspond well with these.^53,54  By contrast, clinical concentrations of midazolam range from 100 to 200 ng/ml (less than 1 μM); the concentrations used here are considerably greater (25 to 400 μM) than those achieved in the clinical setting.⁵⁵  Our cell viability data at nM concentration range did not show any effects of midazolam on either cell line (Supplemental Digital Content, fig. S1, https://links.lww.com/ALN/B760), indicating that midazolam only stimulated apoptosis at concentrations higher than those achieved clinically. These data somewhat highlight the differences between the growth-promoting effect of dexmedetomidine and the lack of effect of midazolam on tumor cell growth at clinically relevant concentrations.

In conclusion, midazolam inhibited cellular proliferation and induced apoptosis through the intrinsic mitochondrial pathway in human lung carcinoma and neuroglioma cells, effects that may be in part mediated by peripheral benzodiazepine receptors. Midazolam also inhibited lung tumor growth in xenograft mice. Dexmedetomidine, however, increased cellular proliferation, migration, and survival through, but not restricted to, pathways activated by α₂-adrenoceptors. Although our study would benefit from deeper investigation into specific receptor activation and cellular signaling, the work reported here furthers the current understanding of the mechanisms of action of dexmedetomidine and midazolam on cancer cells.

The authors sincerely thank Prof. Zhongcong Xie, M.D, Ph.D. (Massachusetts General Hospital, Boston, Massachusetts), for his kind gift of the H4 cell line used in these studies.

The project was supported by a grant from the British Oxygen Company Chair (to Prof. Ma), Royal College of Anaesthetists (London, United Kingdom); European Cooperation in Science and Technology Action (CA15204; Brussels, Belgium; to Prof. Ma); and grant 81600962 (to Dr. Chunyan Wang) from the National Natural Science Foundation (Beijing, China). Dr. Chunyan Wang also received a scholarship (2016041) from the China Scholarship Committee (Beijing, China). Tanweer Datoo received a president’s Ph.D. scholarship from Imperial College London (United Kingdom).

The authors declare no competing interests.