All experiments were performed in adherence with the guidelines for the care and use of laboratory animals (13) and were approved by the Fourth Military Medical University Committee on Animal Care (Xi'an, China). Animal treatments were performed according to the ethical guidelines of the International Association for the Study of Pain (14) for the investigation of experimental pain in conscious animals. Male Sprague Dawley rats (n=24; 7 weeks, 150–220 g) were obtained from the animal research center at the Fourth Military Medical University. The rats were raised in individual standard cages and maintained on a 12 h light/dark cycle (lights on at 07:00 a.m.) at 22°C. Testing was performed during the light cycle. Food and water were available ad libitum. Each animal was restrainer-trained and habituated to the test environment for 1 week prior to testing.

Alfentanil (cat. no. 20151009; Hubei Gedian Humanwell Pharmaceutical Co., Ltd., Wuhan, China) was suspended in propylene glycol (XiLu Chemical Co., Ltd., Guangzhou, China) and diluted with propofol (10 mg/ml; cat. no. p028104; diprivan; AstraZeneca, London, UK) to obtain the desired concentrations of the drug. Rabbit p44/42 MAPK (ERK1/2; 137F5) antibody (cat. no. 4695; 1:1,000), rabbit phospho-p44/42 MAPK [phosphorylated ERK1/2 (pERK); Thr202/Tyr204] antibody (cat. no. 4376; 1:1,000) and rabbit c-fos antibody (cat. no. 2250; 1:1,000) were purchased from Cell Signaling Technology, Inc. (Danvers, MA, USA). Unless otherwise stated, all other chemicals were purchased from Sigma-Aldrich (Merck Millipore, Darmstadt, Germany).

The paw formalin test is a well-characterized experimental model of nociception and was performed as described previously (15). Following intraplantar injection of formalin solution (5%; 50 µl) into the right hind paw, 24 rats (n=6 per group) were placed in individual clear plastic cages (22×12×12 cm). Before the start of the experiments, the animals were acclimatized to the laboratory environment for at least one week. Flinching behavior was considered to be an expression of nociception. Time courses of antinociceptive responses resulting from the administration of different drugs were constructed by plotting the mean number of flinches as a function of time. The typical time course of the response to formalin is biphasic, with an early and short-lasting first phase followed, after a quiescent period, by a second, prolonged (tonic) phase. While phase I is considered to reflect acute nociceptive pain due to a direct stimulation of the nerve by formalin, phase II is attributed to the combination of ongoing inflammatory-associated afferent input from peripheral tissue and functional changes in the spinal dorsal horn (central sensitization) (16). Flinching was defined as rapid and brief withdrawal or flexing of the injected paw, lifting, licking and rubbing behavior. The time-response data were presented as the total number of flinches. To determine the ED50 values of each drug, the number of flinches was converted to the percentage of maximum possible effect (% MPE) according to the formula:

MPE (%)=100-(Sum of flinching count with drug/Sum of control flinching count)x100.

The up-and-down method was used to determine the median effective analgesic dosages of alfentanil and propofol administered either alone, or in combination. Certain groups of animals (n=6 per group) were treated intravenously with various concentrations of either alfentanil (5, 10, 20 and 40 mg/kg) or propofol (2.5, 5, 10 and 20 mg/kg) 10 min prior to formalin injection. The alfentanil-propofol combination was also administered intravenously at increasing doses (0.5, 1, 2 and 4 mg/kg). Control animals were given saline and were interspersed concurrently with the drug-treated animals. The volumes administered were 4 ml/kg. Rats in all groups were observed for changes in behavior that may have been induced by the treatments. In the present study, the initial dosages of alfentanil and propofol were determined according to preliminary experiments and literature values (7,17). Dose-response curves were constructed using least-squares linear regression using the Pharm Tools Pro software (version no. 1.1.27; The McCary Group, Inc., Lehigh Valley, PA, USA). Following the assessment of antinociception for each drug, experimental median effective dose (ED₅₀) values for each drug were determined. Subsequently, a dose-response curve was obtained by concurrent delivery of alfentanil and propofol in a fixed-ratio mixture (1:150), which was based on the ED₅₀ values of each individual agent. The ratio of the two drugs was obtained from our previous study (18).

To construct the experimental antinociceptive effect-dose curve, each group of rats received one of the following doses of the drug combination: Alfentanil ED₅₀ (25.3 µg/kg)+propofol ED₅₀ (8.7 mg/kg); alfentanil ED₅₀/2 (12.65 µg/kg)+propofol ED₅₀/2 (4.35 mg/kg); alfentanil ED₅₀/4 (6.33 µg/kg)+propofol ED₅₀/4 (2.18 mg/kg); alfentanil ED₅₀/8 (3.17 µg/kg)+propofol ED₅₀/8 (1.09 mg/kg) or alfentanil ED₅₀/16 (1.59 µg/kg)+propofol ED₅₀/16 (0.55 mg/kg). The experimental ED₅₀ value for the alfentanil-propofol combination was calculated from this curve. For drug combinations, experimental ED₅₀ (ED_{50 mix}) and its associated 95% confidence intervals (CIs) were determined by linear regression analysis of the log dose-response curve and compared with a theoretical additive ED₅₀ (ED_{50 add}) obtained from the calculation: ED_{50 add}=fxED_{50 alfentanil} + (1-f)xED_{50 propofol}, where f denotes a fraction of the corresponding ED₅₀ in drug mixture (in the present study, f=0.5). In this equation (17,18), ED_{50 add} is the total dose, and the variance of ED_{50 add} was calculated as follows: Var ED_{50 add}=f² × Var ED_{50 alfentanil}+ (1 -f)²x Var ED_{50 propofol}.

From these variances, CIs were calculated and resolved according to the ratio of the individual drug in the combination. When the drug combination gives an experimental ED₅₀ not statistically different from the theoretical calculated ED₅₀, the combination has an additive effect. If the mixture ED₅₀ lies on the theoretical additive ED₅₀ line, or the confidence intervals overlap that line, then the mixture is considered to be additive. If the mixture ED₅₀ lies below the theoretical additive ED₅₀ line and the confidence intervals do not overlap that line, the mixture is considered to be synergistic. The theoretical and experimental ED₅₀ values of the combinations were also contrasted by calculating the interaction index (γ). This was calculated as ED_{50 mix}/ED_{50 add}. If the value is ~1, the interaction is additive. Values <1 are an indication of the magnitude of supra-additive or synergistic interactions, and values >1 correspond to sub-additive or antagonistic interactions (19–22).

The behavioral test was performed on the following 5 groups of rats following formalin injection (5%; 50 µl; n=6 per group): Group FA (rats received formalin and an equal volume of saline vehicle); group Alf (rats received formalin and alfentanil at dose of 25.3 µg/kg); group Pro (rats received formalin and propofol at dose of 8.7 mg/kg); group Alf + Pro (rats received formalin, alfentanil-propofol combination at dose of 2.51 mg/kg); control group rats received an equal volume of saline vehicle only. The dosage of alfentanil or propofol alone was ED₅₀. The dosages of combination of alfentanil and propofol were ED₅.

To evaluate pERK protein levels, rats were sacrificed 30 min after drug treatment. For the measurement of c-fos protein expression, rats were sacrificed 60 min after drug treatment. Rats were anesthetized intraperitoneally with 40 mg/kg sodium pentobarbital, and perfused with fresh 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Following decapitation, after a brief 20 sec exposure to CO₂, rat spinal cord L5-L6 segments were excised, spinal meninges removed and tissue dissociated mechanically, and stored at −80°C until detection. Tissue samples were homogenized by using a Polytron handheld homogenizer (Kinematica, Luzern, Switzerland) in lysis buffer (50 mM Tris-HCl; 150 mM NaCl; 2 mM EDTA; 10% glycerol; 1% Triton X-100; 1% Igepal CA-630; 1% sodium deoxycholic acid). After 30 min incubation, homogenates were centrifuged at 12,000 × g for 15 min at 4°C. Supernatants were then collected and stored at −80°C until used. Total protein content was determined using a Bradford kit from Bio-Rad Laboratories, Inc. (Hercules, CA, USA).

Western blot assay was performed as previously described (21). Aliquots of total protein samples (50 µg) were separated on a 10% SDS-PAGE gel and transferred onto a nitrocellulose membrane. The filter membranes were blocked with 5% non-fat milk for 1 h at room temperature and incubated with the primary antibody specific for pERK1/2 or total ERK1/2 (1:1,000 dilution) and c-Fos (1:1,000 dilution). The membrane was washed with 0.05% TBS Tween buffer and incubated for 1 h with the secondary antibody conjugated with horseradish peroxidase (Goat anti-horseradish peroxidase; cat. no. 123-005-021, 1:1,000; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) for 1 h at room temperature and visualized in ECL solution (cat. no. 1705060; Bio-Rad Laboratories, Inc., Hercules, CA, USA). for 1 min, followed by film exposure for 1–10 min. The loading and blotting of equal amounts of proteins were verified by re-probing the membrane with antibody against β-actin (1:1,000; cat. no. sc-47778, Santa Cruz Biotechnology, Inc., Dallas, TX, USA). The intensity of each immunoblot assay band was quantified using a VersaDoc Imaging System (Bio-Rad Laboratories, Inc.). The experiments were repeated twice. Quantification of immunoreactivity corresponding to the total and phosphorylated bands was performed by densitometric analysis using Multi Gauge Version 3.0 (Fujifilm, Tokyo, Japan).

Results were presented as the mean ± standard error of the mean or as ED₅₀ values with 95% CIs. The statistical significance of dose-responses was determined by one-way analysis of variance followed by the Tukey's post hoc test. Isobolographic calculations were performed by using the Pharm Tools Pro software (version no. 1.1.27, The McCary Group, Inc.). Statistical analysis of the isobolograms was performed as previously described (22) and differences between experimental and theoretical values were assessed by Student's unpaired t-test. P<0.05 was considered to indicate a statistically significant difference.

Plantar injection of formalin produces nociceptive behavior, including flinches of the paw. Flinching was defined as rapid and brief withdrawal or flexing of the injected paw, lifting, licking and rubbing behavior. In the present study, the number of pain responses in 5 min intervals during phase I, for 20 min, and 10 min intervals during phase II, for 40 min, following formalin injection was recorded. Saline-treated control rats exhibited discrete biphasic behavioral responses consisting of an early short-lasting response (phase I, 0–10 min post-injection), followed by a late, prolonged response (phase II, ~16–60 min post-injection). The duration of licking, lifting and rubbing were considered to be nociceptive behaviors in the formalin model. The mean number of flinches peaked around 0–10 min and 30–50 min after formalin intraplantar injection, which was followed by a gradual decline in all groups (Fig. 1A). Nociceptive behavior, the mean number of flinches, between saline and drug-treated groups was compared. There were no significant differences between any of the groups during phase II of the behavioral response. However, the amount of licking and lifting behavior was reduced in the alfentanil or propofol alone groups, and alfentanil-propofol combination group for the first 5 min after formalin injection, compared with the saline group during phase I (P<0.05; Fig. 1A). The total numbers of flinches during phase I and phase II following formalin injection in the saline treatment group was 121.4±8.8 and 100.5±12.1, respectively. However, compared with the saline treatment group, the total number of flinches was significantly decreased in groups Alf (56.1±5.3), Pro (50.3±5.0) and Alf+Pro (33.2±4.6) during phase I (P<0.05; Fig. 1B). There were no significant differences between any of the groups during phase II.

In phase I, alfentanil, propofol and alfentanil-propofol combination groups led to a dose-dependent antinociceptive effect. Linear regressions for alfentanil (Y=186.0X-24.7; R²=0.9861), propofol (Y=153.7X-10.92; R²=0.9728) and co-administration (Y=91.97X+8.55; R²=0.9962) were calculated by plotting MPE (%) against the log dose (Fig. 2). The ED₅₀ of individual administration of alfentanil and propofol were 25.3±2.21 µg/kg and 8.7±1.344 mg/kg, respectively (Table I). The isobologram was constructed by connecting the ED₅₀ of alfentanil on the abscissa with the ED₅₀ of propofol on the ordinate to obtain the additive line (Fig. 3). For the drug combination, the ED_{50 mix} and the 95% CI of the mixture were computed by linear regression of the log dose-response curve. The ED_{50 mix} and ED_{50 add} were plotted in the isobologram (Fig. 3). The fixed drug-dose ratio based on mass quantity for alfentanil and propofol is 1:344. The total ED_{50 mix} for the alfentanil-propofol combination is 2.51±0.56 mg/kg (Table I), representing 7.27 µg/kg alfentanil and 2.50 mg/kg propofol. By isobolographic analysis, the ED_{50 add}=4.36 mg/kg [(0.5xED₅₀ alfentanil) + (0.5xED₅₀ propofol)] (Table I), representing 12.64 µg/kg alfentanil plus 4.35 mg/kg propofol. The ED_{50 mix} was <ED_{50 add} (Table I). The γ value was 0.57, which suggests a synergistic interaction between alfentanil and propofol during phase I of the formalin test (Table I).

The expression of total and phosphorylated-ERK (Fig. 4), and c-fos (Fig. 5) in control and drug-treated groups was detected. Compared with the β-actin bands, the ratios of the optical density of total ERK bands in all groups were not significantly different. To investigate the expression level of the two proteins, the ratio of the optical density of each band for phosphorylated and total proteins were detected, and values were normalized as a percentage of the control group. The effect of drug treatment on total and pERK expression was determined in spinal cord L4-L5 segments of the lumbar enlargement, areas that are involved in the transmission of nociceptive inputs and the control of sympathetic outflow (23). In the present study, two bands, p42 and p44, represented ERK and were analyzed. Concerning ERK activation, there was no significant change in the expression of total-ERK following formalin addition (Fig. 4). However, formalin induced the activation of pERK, which was significantly reduced when subsequently treated with alfentanil (P=0.02) or propofol (P=0.03) alone, or combination treatment (P=0.01), compared with the formalin only group (FA). Furthermore, formalin-induced activation was significantly reduced by combination treatment (Pro + Alf) compared with those treated with propofol (P=0.01) or alfentanil (P=0.03) only groups (Fig. 4). The current study subsequently investigated whether propofol and alfentanil had any effect on basal levels of c-fos expression. As illustrated in Fig. 5, there was a clear increase in the expression of c-fos following injection of formalin, highlighted in the FA group. This increased the level of c-fos expression was then significantly reduced when treated with alfentanil (P=0.02), propofol (P=0.02) or combination treatment (P=0.03) compared with the FA group. Notably, the Alf+Pro group also exhibited significantly decreased c-fos expression when compared with treatment with alfentanil (P=0.01) or propofol (P=0.02) alone (Fig. 5).