The Inhibitory Effects of Thiopental, Midazolam, and Ketamine on Human Neutrophil Functions

Abstract

We investigated the effect of thiopental, midazolam, and ketamine (at clinically relevant concentrations and at 0.1 and 10 times these concentrations) on several aspects of human neutrophil functions. The three intravenous (IV) anesthetics significantly decreased chemotaxis, phagocytosis, and reactive oxygen species (ROS) (O_2-, H₂ O₂, OH) production of neutrophils in a dose-dependent manner. At clinically relevant concentrations, thiopental and midazolam significantly depressed these neutrophil functions. However, ketamine at the clinical plasma concentration did not impair chemotaxis or ROS production, except phagocytosis. In contrast, the three anesthetics had no effect on the levels of ROS production by a cell-free ROS generating system. In addition, intracellular calcium concentrations in neutrophils stimulated by N-formyl-L-methionyl-L-leucil-L-phenylalanine were dose-dependently decreased in the presence of each of the three anesthetics. The suppression of an increase in intracellular calcium concentrations may be responsible for the inhibition of neutrophil functions by the IV anesthetics. Implications: Neutrophils play an important role in the antibacterial host defense system and autotissue injury. We found that thiopental and midazolam (but not ketamine), at clinically relevant concentrations, impaired the neutrophil functions.

The immune function of neutrophils has been called a double-edged sword. Neutrophils play a central role in the antibacterial host defense mechanism as a component of nonspecific cell-mediated immunity ^[1]. The neutrophil response to microbial invasion may be classified into chemotaxis, adherence, phagocytosis, and intracellular killing (consisting of formation of phagosome, degranulation of peroxidase and hydrolases, and microbial death and digestion) ^[1]. On the other hand, neutrophils may also play a pivotal role in the pathogenesis of autotissue injury, leading to organ failure such as acute respiratory distress syndrome ^[2]. The inadvertent excessive release of reactive oxygen species (ROS) from neutrophils that accumulate in the organ by chemotactic factors into the surrounding milieu contributes to this process.

Impairment of neutrophil function is therefore thought to act advantageously or disadvantageously. A disorder of any neutrophil function may allow a bacterial infection to develop. The inhibitory effects of some anesthetics, including intravenous (IV) anesthetics, on neutrophil functions have been well documented ^[3]. The use of these anesthetics may result not only in perioperative immunosuppression, but if used to sedate critically ill patients in an intensive care unit (ICU), may potentially further compromise an already depressed host defense mechanism. In contrast, suppression of neutrophil functions by the anesthetics may be beneficial to the organ dysfunction induced by autotissue injury.

Thiopental, midazolam, and ketamine have been used extensively for anesthesia and/or sedation. Many researchers have independently shown that at concentrations higher than those clinically observed, these IV anesthetics inhibit human neutrophil functions, including chemotaxis ^[4-6], phagocytosis ^[7,8], and ROS generation ^[9,10]. However, there are some conflicting data on the effects at clinically relevant concentrations ^[4-15]. The types of oxygen metabolites (superoxide (O (_2-)), hydrogen peroxide (H₂ O₂), and hydroxyl radical (OH) have not been specified in previous studies using chemiluminescence. H₂ O₂ and OH, in particular, are considered to be important as major toxic candidates because of their most potent oxidative activity to kill bacteria or to cause injury to the host tissues ^[16]. Concerning the neutrophil functions, limited comparative data of these IV anesthetics are available. In the current study, therefore, we compared the effects of thiopental, midazolam, and ketamine on neutrophil chemotaxis, phagocytosis, and production of O_2-, H₂ O₂, and OH. We also examined whether these anesthetics can scavenge an excess of the ROS generated.

Methods

After receiving institutional approval and informed consent, neutrophils were isolated from heparinized venous blood from 12 healthy volunteers according to methods previously described ^[17]. Briefly, after centrifugation of the blood over a Ficoll-Hypaque gradient, the plasma-containing upper layer, mononuclear cell layer, and remaining cell pellet were removed separately. The plasma was freed of platelets by using centrifugation. The cell pellet, containing neutrophils and erythrocytes, was washed with saline solution and resuspended in plasma containing dextran 170 at a final concentration of 1%. The neutrophils were recovered after sedimentation at unit gravity, and the few contaminating erythrocytes were lysed by treatment of the preparation with 0.876% NH₄ Cl. The neutrophils were then resuspended in media appropriate for their subsequent use: RPMI 1640 medium for chemotaxis assay, Krebs Ringer phosphate (KRP) buffer for phagocytosis assay, KRP containing glucose (5 mM) for OH generation, and KRP buffer containing glucose (5 mM) and gelatin (1 mg/mL) for the assay of O_2- and H₂ O₂ generation.

Neutrophil viability after incubation with the agents was determined by means of the trypan blue exclusion test, and phagocytic function was measured by zymosan-induced stimulation of¹⁴ C-inulin uptake ^[18]. When more than 2% of the neutrophils were stained with trypan blue or when the cells absorbed less than 600 dpm¹⁴ C-inulin per milligram of protein, their function was considered to have been impaired and the results were discarded.

None of the IV anesthetic (i.e., thiopental, midazolam, and ketamine) contained any preservative, and all were diluted in KRP. The IV anesthetics (0.1 or 0.2 mL except chemotaxis assay) were placed in the measuring cuvette to a total volume of 1 or 2 mL, giving the following final concentrations of the IV anesthetics: thiopental (Ravonal[registered sign]; Tanabe, Osaka, Japan) and ketamine (Ketalar[registered sign]; Sankyo, Tokyo, Japan) at 0, 3, 30, and 300 micro g/mL, and midazolam (Dormicum[registered sign]; Yamanouchi, Tokyo, Japan) at 0, 0.15, 1.5, and 15 micro g/mL. However, for chemotaxis assay, 20 micro L of each anesthetic was added to the upper compartment of a Boyden chamber to a total volume of 0.2 mL. These concentrations of anesthetics corresponded to 0.1, 1, and 10 times clinical plasma concentrations.

Neutrophil chemotaxis was determined using a modified Boyden chamber as previously described ^[19]. A glass tube (4.5 mm in diameter, 12.0 mm long) was used as the upper compartment of the chamber (Figure 1). The bottom surface was attached to a nonsiliconized polycarbonate filter (3.0-micro m pore, Bio-Rad Lab., Tokyo, Japan), using Eukit, a mounting reagent (O. Kindler, Freiburg, Germany). A vial 4.5 mm in diameter and 4.8 mm in length was used as a lower compartment of Boyden chamber. The neutrophils (0.2 x 10⁶ cells/0.2 mL of total volume) containing 70 micro L of neutrophil medium (RPMI 1640 medium) with each concentration of thiopental, midazolam, or ketamine were placed into the upper compartment of the chamber, and the concentrations of 10^-7 M N-formyl-L-methionyl-L-leucyl-L-phenylalanine (FMLP), a potent synthetic chemoattractive peptide, were placed into the lower compartment (total volume 0.2 mL). The chamber was incubated at 37[degree sign]C for 45 min in an atmosphere of 5% CO₂. The number of cells reaching the bottom surface of the filter was expressed as the average number of cells per field after counting five fields for each filter.

Figure 1: Schematic structure of a Boyden chamber. The Boyden chamber is an apparatus to measure chemotaxis of neutrophils in vitro. This device fundamentally consists of an upper (a) and a lower (c) compartment; neutrophils and chemoattractants are applied to the upper and lower compartments, respectively. A filter (b) with many micropores separates the two compartments.

Emulsions of paraffin oil containing Oil Red O were prepared as previously described ^[20], except that the lipopolysaccharide solution was replaced with normal human serum. The emulsion was incubated with an equal volume of normal human serum at 37[degree sign]C for 30 min for opsonization. Neutrophils (2 x 10⁷ cells/0.9 mL KRP) that had been preincubated for 5 min with each concentration of thiopental, midazolam, or ketamine were added to 0.1 mL of the opsonized emulsion. The mixture was incubated for 5 min at 37[degree sign]C, and then 9 mL ice-cold KRP was added to the solution to stop the reaction. The cells were washed three times with ice-cold KRP to remove the paraffin oil droplets that had not been ingested. Paraffin oil containing Oil Red O was extracted from the cells by the method described by Bligh and Dyer ^[21], using chloroform and methanol (1:2 vol/vol). The optical density of the chloroform layer was determined at a wavelength of 525 nm. The mean optical density of Oil Red O extracted from 2 x 10⁷ neutrophils incubated with opsonized paraffin oil droplets was 0.025 +/- 0.0019 (mean +/- SD of five experiments), and microscopic examination revealed that most neutrophils were heavily loaded with oil droplets. On the other hand, when nonopsonized paraffin oil droplets were incubated with neutrophils, the optical density was less than 0.005, and under the microscope, only a few neutrophils were seen to be loaded. These findings confirmed that most of the extracted Oil Red O represented droplets ingested by the neutrophils.

For O_2- formation, 1 x 10⁶ neutrophils were prein-cubated at 37[degree sign]C for 10 min with 1 mg/mL opsonized zymosan (Sigma Chemical Company, St. Louis, MO), and various concentrations of thiopental, midazolam, or ketamine, and then 0.1 mM ferricytochrome-c (type III of the same lot, more than 95% of oxidized form; Sigma) were added. The neutrophils were incubated for another 30 min. Immediately after sedimentation of the neutrophils and opsonized zymosan by centrifugation, 0.1 mL of the supernatant was assayed for reduced cytochrome c by measuring the absorbance at 550 nm in 2 mL of 100 mM potassium phosphate buffer (pH 7.8) containing 0.1 mM EDTA (pH 7.8) ^[17,19]. The results were converted to nmol of reduced cytochrome c, using Delta E_550nm = 2.1 x 10⁴ M/cm ^[22].

The H₂ O₂ generation was measured by quantifying the decrease in fluorescence intensity of scopoletin (Sigma) due to its peroxidase-mediated oxidation by H₂ O₂^[17,19,22]. After incubation of 2.5 x 10⁶ neutrophils for 10 min at room temperature in KRP containing 5 mM glucose and 0.1 mg/mL gelatin in the presence of each concentration of the IV anesthetics and 1 mg/mL opsonized zymosan, 0.1 mL of 50 mM scopoletin in KRP and 0.05 mL of 1 mg/mL horseradish peroxidase (type II; Sigma) in phosphate-buffered saline (PBS) were added. The rate of H₂ O₂ plus peroxidase-induced rate of decrease in fluorescence intensity of the scopoletin within 30 min was quantified using a fluorescence spectrophotometer (Hitachi). The H₂ O₂ concentration was calculated assuming that 1 M H₂ O₂ oxidizes 1 M scopoletin ^[22].

The OH generation was determined by the amount of ethylene gas formed from alpha-keto-methiolbutyric acid (KMB) (Sigma) plus the OH generated by neutrophils ^[17,19]. Neutrophils (2 x 10⁶) in KRP containing glucose were preincubated with 1 mM KMB and each concentration of the IV anesthetics in a stoppered tube and gently mixed in a shaker bath at 37[degree sign]C for 5 min. Opsonized zymosan was then added, and the cells were incubated for 10 min. Aliquots of gas in the tube were sampled using a gas-tight syringe, and the ethylene content was determined with a gas chromatograph (Hitachi), using nitrogen gas as a carrier and Porapak Q column (3 x 1 m; Hitachi) with a temperature controller set at 130[degree sign]C (temperature rate 3-5[degree sign]C/min). The total amount of ethylene formed during 10, 20, and 30 min represented the OH value.

All ROS were also generated by the reaction of hypoxanthine with xanthine oxidase. Instead of neutrophils and opsonized zymosan, 0.1 mM hypoxanthine, 1.25 mM EDTA, and each of the IV anesthetics were mixed in a total volume of 2 mL (125 mM phosphate buffer). This reaction mixture was incubated for 2 min at 37[degree sign]C. Approximately 0.006 U/mL of dialysed xanthine oxidase was added to the mixture to start the reaction. The O_2-, H₂ O₂, and OH generated in the cell-free system was measured by the same method used in the neutrophils system.

Fura-loaded, FMLP-stimulated [Ca²⁺]_i increase was assessed according to the method described in the literature with some modifications ^[23]. Fura-2 A/M, at a final concentration of 1 micro M, was added to the purified neutrophil suspension in 0.1 mM CaCl₂ containing KRP and incubated for 30 min at 37[degree sign]C in a cuvette with mild shaking. After incubation, the fura-2-loaded cells were washed twice with KRP and resuspended in KRP at a concentration of 10⁷ cells/mL. Thiopental, midazolam, or ketamine were then added at the indicated concentrations and incubated for a further 5 min (total volume 2 mL). Fifteen milliliters of 10 (^-6) M FMLP was added to 1.5 mL of this mixture to measure fura-2 fluorescence using a Hitachi F-4000 fluorescence spectrophotometer, with the excitation wavelength alternated every 4 s from 340 nm to 380 nm, with the emission wavelength set at 510 nm. The neutrophil suspension was kept at 37[degree sign]C with constant stirring throughout the measurement. Calcium concentrations were determined from the ratio of fura-2 fluorescence intensities at excitations of 340 nm and 380 nm ^[23]. The data were processed with a computer fitted to the Hitachi F-4000 fluorescence spectrophotometer, and estimated [Ca²⁺]_i levels were recorded sequentially. The calcium concentration was expressed by the following formula: Equation 1 where R is the ratio of F340 nm to F380 nm, Kd is the K value of Ca²⁺ and fura-2, R_max is sufficient Fura-2 bound to Ca²⁺ in the presence of 10% Triton X-100 (30 micro L), R_min is the minimal amount of Fura-2 bound to Ca²⁺ in the presence of 100 mM EGTA (pH 9.3), and b is the ratio of F380 nm in the absence of Ca²⁺ to F380 nm in the presence of Ca²⁺.

The results are expressed as the mean +/- SD (n = 12: the number of volunteers). Triplicate assays were performed for all experiments (chemotaxis, phagocytosis, ROS, and Ca²⁺) using neutrophils isolated from each volunteer (n = 1). Data were analyzed for statistical significance by using repeated-measures analysis of variance. The sample size of the current study was sufficient to detect large differences (d = [micro 1-micro 2]/sigma = 0.7-1.0) in ROS production at a significance level of 0.05, although the power of the study is relatively weak (power 1 -beta = 0.5-0.7).

Results

Thiopental, midazolam, and ketamine exerted a dose-dependent inhibitory effect on neutrophil chemotaxis and phagocytosis (Figure 2). The impairment of phagocytosis by thiopental seems to be the greatest. At the clinically relevant concentration (30 micro g/mL = 109 micro mol/m) ketamine failed to depress chemotaxis.

Figure 2: Effects of thiopental (a), midazolam (b), and ketamine (c) on neutrophils' chemotaxis and phagocytosis. Data (mean +/- SD are expressed as a percentage of the control (in the absence of intravenous anesthetics). Closed and dotted columns indicate chemotaxis and phagocytosis, respectively. # = clinical plasma concentration of each anesthetic. *P < 0.05 versus control.

The O_2-, H₂ O₂, and OH generated by human neutrophils were decreased in a dose-dependent manner in the presence of the IV anesthetics (Figure 3). The inhibitory effect of thiopental was greater than that of the other two anesthetics. Ketamine had no effect on the ROS production by neutrophils at the clinically relevant concentration but inhibited it at higher concentrations (Figure 3). In contrast, none of the ROS generated in the xanthine-xanthine oxidase system was impaired by these anesthetics at any concentration (Figure 4). These findings indicate that thiopental, midazolam, or ketamine did not scavenge the ROS generated, but rather inhibited the ability of neutrophils to produce ROS.

Figure 3: Effects of thiopental (a), midazolam (b), and ketamine (c) on reactive oxygen species produced by neutrophils. Data (mean +/- SD) are expressed as a percentage of the control (in the absence of intravenous anesthetics). Closed, dotted, and striped columns indicate O_2-, H₂ O₂, and OH, respectively. # = clinical plasma concentration of each anesthetic. *P < 0.05 versus control.

Figure 4: Effects of thiopental (a), midazolam (b), and ketamine (c) on reactive oxygen species produced by xanthine-xanthine oxidase system. Data (mean +/- SD) are expressed as a percentage of the control (in the absence of intravenous anesthetics). Closed, dotted, and striped columns indicate O_2-, H₂ O₂, and OH, respectively. # = clinical plasma concentration of each anesthetic. *P < 0.05 versus control.

Fura-loaded and FMLP-stimulated [Ca²⁺]_i in the neutrophils decreased in a dose-dependent fashion in the presence of the anesthetics (Figure 5).

Figure 5: Effect of thiopental (a), midazolam (b), and ketamine (c) on intracellular calcium ion concentrations in neutrophils stimulated by N-formyl-L-methionyl-L-leucil-L-phenylalanine (FMLP). Data (mean +/- SD) are expressed as a percentage of the control (in the absence of intravenous anesthetics). # = clinical plasma concentration of each anesthetic. *P < 0.05 versus control.

Discussion

We have shown that at clinically relevant concentrations, the IV anesthetics suppressed neutrophil functions. The impairment could affect a patient's ability to fight infection. This may be more serious if the anesthetics are administered continuously during the period of infection or when infection is likely to occur, particularly in the ICU. Long-term treatment with barbiturates has been associated with an increased susceptibility to certain infections. Thiopental has been shown to induce a greater incidence of pneumonia in ventilated head-trauma patients ^[24]. The ROS generated by neutrophils accumulated in various organs in response to chemotaxins is thought to play a pivotal role in the pathogenesis of multiple organ failure induced by sepsis or endotoxemia ^[2]. However, limited data on the attenuative effect of thiopental, midazolam, and ketamine on the acute organ injury are available. Thus, further studies are required to verify this beneficial effect. We used neutrophils obtained from adult healthy volunteers in the current study. The use of the phagocytes from patients with sepsis, acute respiratory distress syndrome or multiple organ failure, in whom the function of neutrophils is probably modulated, may elicit different results. Psychotomimetic side effects of ketamine are sufficient to nullify the beneficial effects of N-methyl-D-aspartate receptor antagonist for cerebral antiischemic therapy. This untoward side effect restricts the use of ketamine in the ICU. Thus, we should have investigated the effects of propofol (instead of ketamine) on neutrophil functions, because use of the drug for sedation in the ICU is increasing and practical.

Conflicting in vitro data on the effects of the anesthetics used in the current study on neutrophil functions have been reported. In consistent with our results, clinical therapeutic concentrations of ketamine ^[7,8] and thiopental ^[8] impaired phagocytosis. Contrary to our findings, Davidson and co-workers ^[11] reported that clinically relevant levels of thiopental, midazolam, and ketamine had minimal effects on neutrophil phagoytosis, as determined by flow cytometry. In their study, whole blood and Staphylococcus aureus were incubated for 60 minutes. Although thiopental inhibited chemotaxis ^[4-6], in agreement with our results, the two research groups indicated failure in depression of chemotaxis ^[12,13]. One group used opsonized zymosan as a chemoattractant ^[12], and the other assessed this function in agarose assay system ^[13]. There is general agreement that thiopental at concentrations higher than those clinically relevant also inhibit ROS generation ^[9-11]. However, thiopental at clinical plasma concentrations decreased chemiluminescence ^[9,10] but failed to reduce respiratory burst ^[11], which represented H₂ O₂ production, as determined by flow cytometry technique. In the latter experiment, whole blood was stimulated by phorbol myristate acetate (PMA). Ketamine at higher concentrations did not affect the ROS production assessed by respiratory burst (H (₂) O₂ production) ^[11] or oxygen consumption (measured by PO₂ decrease after stimulation with FMLP or PMA) ^[14]. These results are contradictory to our observation concerning the ROS generation. Midazolam at clinically relevant concentrations depressed FMLP-induced chemiluminescence ^[15]. Our results support these data. However, other investigators ^[11] demonstrated that PMA-induced respiratory burst activity was unchanged in the presence of midazolam at 100 times the plasma concentration. Neutrophil functions are clinically influenced by multiple factors, which are present in the circulation or tissue, including stress, hormonal changes, cytokines, underlying disease, age, previous medications, and blood transfusion. Because our experimental protocol excluded these factors, the current study enabled us to elicit the essential effects of the IV anesthetics. Our study is the first to specify the effects of the IV anesthetics on OH production. These may be advantages.

Nicotinamide adenine dinucleotide phosphate oxidase present in the cell membrane of neutrophils is the enzyme responsible for the production of O_2-^[1]. The two pathways by which extracellular stimuli are transmitted to activate nicatinamide adenine dinucleotide phosphate oxidase are proposed: protein kinase C (PKC) activation and increased [Ca²⁺]_i (Figure 6) ^[1]. Immediately after FMLP is bound to the receptor, the signal is transmitted to phospholipase C (PLC) through GTP-binding protein, with the resultant activation of PKC and inositol triphosphate, which in turn mobilize [Ca²⁺]_i from the intracellular calcium storage site (endoplasmic reticulum) ^[1]. We have shown that thiopental and midazolam at clinically relevant concentrations (and high concentration of ketamine) suppressed the increase in [Ca²⁺]_i of neutrophils stimulated with FMLP, whereas clinical plasma concentrations of ketamine failed to do so. The inhibition of FMLP-induced [Ca²⁺]_i increase by the IV anesthetics may be due to decreased PLC activation. At clinically relevant concentrations, thiopental and midazolam but not ketamine reduced the ROS generation from neutrophils. Although the precise mechanism of the inhibitory effect of the IV anesthetics on ROS production remains to be elucidated, suppression of the increase in [Ca²⁺]_i may contribute to the inhibition of the neutrophil function. Thiopental, midazolam, and ketamine, which are associated with high lipid solubility, tend to accumulate in lipid membranes, altering their physical properties. The resultant changes in membrane-bound enzymes may be a mechanism of action of the IV anesthetics. We previously reported that barbiturates, inhibit PKC activation ^[25]. Thiopental may impair the ROS production, partly through this mechanism.

Figure 6: Scheme of mechanisms of FMLP-induced Ca²⁺ elevation and O₂ production in neutrophils. FMLP=N-formyl-L-methionyl-L-leucil-L-phenylalanine, PMA=phorbol myristate acetate, G=guanosine triphosphate-binding protein, PLC=phospholipase C,PIP₂=phosphatidylinositol bisphosphate,^I P₃=inositol triphosphate, DG=diacylglycerol.

In conclusion, we have shown that thiopental and midazolam at clinically relevant concentrations inhibit the chemotaxis, phagocytosis, and ROS production of neutrophils. At clinically plasma concentrations, ketamine impaired phagocytosis but not chemotaxis or ROS production. The reduction of these neutrophil functions by the anesthetics may be due to a suppression of the increase in [Ca²⁺]_i.