Tramadol hydrochloride injection (THi) was purchased from Medochemie Ltd (Limassol, Cyprus), and tramadol hydrochloride tablets (THt) were from Wellso Pharmaceutical Co., Ltd. (Beijing, China). Solvents used were HPLC grade and purchased from Tedia Company, Inc. (Fairfield, OH, USA). Milli-Q water (EMD Millipore, Billerica, MA, USA) was used throughout. Chemicals were all reagent grade.

Zebrafish (Danio rerio) were obtained from Wei Tong Lihua BioModel Company (Beijing, China) and raised to maturity in a stock tank without filters at a temperature of 28°C under cycles of 14 h light and 10 h dark. Their daily care conformed a guide for the laboratory use of zebrafish (24).

Experimental fish were randomly selected from the stock tank, and maintained in individual beakers with covers for 2 days. IM administration of TH_i was performed using a microinjector and self-made equipment with a superficial hole to hold the fish. In this equipment, wet absorbent cotton was fixed to keep the body of the fish wet, above which several rubber bands were used to strengthen the hold. Sites for IM injection were the flanks, usually midpoint between the dorsal fin and the lateral line. A volume of drug or saline was injected into treatment group or control group, respectively, and no treatment was administered to the vehicle group. Fish were placed back into their beakers. For the oral administration of tramadol, zebrafish were pre-exposed for a certain time to a solution of tramadol, with TH_t directly dissolved into the tank. Experiments were performed in accordance with the guidelines published in the NIH Guide for the Care and Use of Laboratory Animals.

Zebrafish were randomly divided into 3 groups (n=6 per group). Fish were IM-injected with 4 µl saline or THi (10 µg/µl), or received no treatment. Tissue samples from brain, eyes, gill, heart, liver, gut and muscle (from the tail of fish, far away from the injection point), were isolated at 0.5 or 1 h after the treatment, and washed with cold 0.9% saline four times, dried with tissue paper, weighed and transferred to glass tubes (1 cm I.D. on ice). The treatment time for every fish was strictly controlled, from injection to tissue isolation.

Tissue samples with 9-fold (w/v) excess saline were homogenized using a microhomogenizer at 2,000 rpm/min for 10 sec and repeated 5 times. A 1 ml aliquot of ethanol was added into the homogenate. The tubes were sealed and shaken in a reciprocal shaker for 2 h, following centrifugation at 12,000 × g at room temperature for 10 min, and the supernatant was transferred to fresh disposable glass tubes. The ethanol extraction procedure was repeated four times. Pooled the supernatant together, then were evaporated under a stream of nitrogen at 40°C, finally the residue was stored at −20°C prior to analysis. The residue was dissolved in η-hexane at the same ratio (w/v), vortexed and centrifuged at 200 × g for 10 min at 4°C prior to ESI-Q-TOF/MS and GC-MS analysis.

The fish were randomly divided into two groups, and received a single IM dose of either 10 or 25 µg/µl THi. Brains (n=7) were collected prior to drug administration (pre-injection, time 0), and at 1, 3, 5, 15, 25, 35, 45, 75, 120 and 240 min after a single dose of 10 µg, or at 1, 2, 4, 5, 7, 10, 20, 30, 40, 55, 95, 155, 180 and 300 min after a single injection of 25 µg. in order to optimize the dose, injection and isolation time. Brain samples were immediately placed on ice, washed with 0.9% saline (pH 9.0) four times, then the supernatant was discarded following washing with 0.9% saline and extracted as aforementioned. A 1 µl aliquot volume of each sample was injected into the GC-MS system. For quantification, calibration curves were created and linearity calculated. A standard calibration curve of tramadol content vs. mean peak area in GC-MS analysis was conducted in the range of 0–10 µg, by adding 0, 0.5, 5, and 10 µg tramadol into the brain samples from untreated vehicle control fish and using the aforementioned extraction procedures.

Primary metabolites, M1 and M2 were prepared from tramadol hydrochloride as previously described (25), with certain modifications. Briefly, 218 mg of platinum oxide hydrate was dissolved in 1 ml methanol, and reduced with hydrogen gas for 5 min. Following centrifugation at 200 × g for 10 min at 4°C, the precipitated platinum black catalyst was washed twice with 0.5 ml water. Then tramadol hydrochloride (20 mg dissolved in 1 ml of distilled water) was immediately added to the catalyst, and the mixture was stirred for 90 h. The reaction mixture was centrifuged at 200 × g for 10 min at 4°C and the supernatant was removed and the residue was washed with 1 ml water. The combined aqueous phases were treated with 0.5 ml 3 mol/l NaOH solution and extracted twice with 2 ml of η-hexane. The hexane phase was evaporated under a stream of nitrogen. The resulting compound was used as the desmethyltramadol reference sample in the GC-MS analysis.

Samples for metabolite analysis were prepared following IM and bath exposure. Fish received a single IM dose of either TH_i (25 µg/µl) or saline, or for bath exposure a solution of 0.06 mg/ml TH_t was added to the tank for 1 h. Brain samples (n=12) were obtained at 0.5, 1 and 2 h after the treatment, which were washed, homogenized and extracted in alkalic conditions as aforementioned, for the reference substance.

ESI-Q-TOF mass spectra were recorded on a Bruker Autoflex III TOF/TOF 200 mass spectrometer (Bruker Corporation, Ettlingen, Germany). The parameters were as follows: Ion mode, positive; source temperature, 200°C; ion spray voltage, 4.5 kV; declustering potential, 30 V; and mass range, 75–800 m/z. A 20 µl aliquot of each sample was analyzed by infusion using a syringe pump for direct injection into ion sources, at 2 µl/min flow rate. MS/MS parameters were: Collision energy, 25 V; and collision gas pressure (N2), 4.7 mPa.

Triple-quadrupol Varian mass spectrometer MS 1200 equipped with a CP-3800 gas chromatograph (Agilent Technologies GmbH, Waldbronn, Germany) was used to detect tramadol and the metabolites. The analytes were separated on a Varian Factor Four VF-5 ms capillary column (30 m × 0.25 mm ID, 0.25 µm df; Agilent Technologies GmbH) with helium 5.0 as the carrier gas (constant flow, 0.2 ml/min for tramadol detection and 1 ml/min for desmethyltramadol detection). The front injector temperature was 260°C. The oven temperature initially set at 100°C for 1 min, was increased to 220°C at 20°C/min, held for 10 min, and then increased to 300°C at 20°C/min, and finally held for 9 min. Injection was performed in splitless mode (10:1; 60 sec delay before opening the splitter). Ionization was performed in electron impact (EI) mode with electron energy of 70 eV. Analytes were measured in full scan mode within the mass range, m/z=50−400. Selected-ion-monitoring (SIM) mode for quantitative analysis from 6 to 40 min was used, m/z 58, 263 for tramadol; m/z 44, 249 for M2; m/z 58, 121, 249 for M1.

In the current study, two types of tramadol, injection (TH_i) and tablets (TH_t), with the same main component, were used. The UV absorption spectra of both types were monitored a single well-defined maximum peak either dissolved in CHCl₃ or water at 271 nm, in the wavelength range of 200–400 nm with a fixed slit width of 5 nm (Fig. 1A). Furthermore, only a single molecular ion peak was detected by ESI-Q-TOF/MS or GC-MS, at m/z 264.2578 (Fig. 1B) and 263 (Fig. 1C), respectively. No structural differences were observed, and no interfering absorbance was detected. Therefore, both types were used as the parallel treatment groups in the subsequent experiments.

Total mass spectra of the brain tissue samples of zebrafish are presented in Fig. 2A. By comparing the tramadol group with the untreated group, an ion peak at m/z 264 in every tissue from tramadol administration group was easily detected (data not shown). Chosen parent peaks for further ionization, seven high-resolution fragment ion peaks were identified, namely at m/z 58.3564, 101.3825, 129.7751, 171.2543, 200.4007, 222.2337 and 264.4308 (Fig. 2B), and the molecular weight differences between the two adjacent fragments were 43, 28, 42, 29, 22 and 42, respectively. Among them, the peak at m/z 58.3564 corresponded to [C₃H₈N], and on this basis, new chemical structures were added one by one according to the differences, and a series of structures were established, [C₆H₁₅N], [C₈H₁₉N], [C₁₀H₂₁NO], [C₁₂H₂₅NO], [C₁₄H₂₅NO], and finally the structure of tramadol [C₁₆H₂₅NO₂] was obtained. The details of the whole process of structure analysis are presented in Table I. Thus, ESI-Q-TOF/MS is considered to be a rapid and adequately sensitive method for the detection of tramadol distribution in tissues.

The compounds of the brain extract in the hexane fraction, with or without tramadol treatment, were performed to optimize the analysis conditions. Comparing the chemical compositions of brain extract from control with tramadol IM administration, the peaks at m/z 263 were only detected in the tramadol group, without any interfering peak from other compounds, within retention time of 13.824 min. It was exactly the tramadol peak, identified by comparison of its mass spectra with that from the software library, which was further confirmed using SIM mode by monitoring for tramadol and [C3H8N]⁺ at m/z 263 and 58, respectively.

Total ion chromatograms of extracts from brain tissues, and enlarged portion of ion chromatograms from other six other tissues are presented in Fig. 3. Using full scan or SIM mode, tramadol peaks were detected in all tissues from the tramadol administration group, at 0.5 or 1 h after injection. The results were highly consistent with that of ESI-Q-TOF/MS. For each tissue, the sample was weighed, prepared by the same process, and dissolved in η-hexane at the same ratio (w/v). Therefore, the same tissue type from different treatment times can be compared. The mean peak area obtained by three repeated detections was substituted into the formula, namely the standard calibration curve of tramadol content vs. the mean peak area: y=0.3669x-0.0184 (R²=0.999, 0≤x≤10 µg) or y=24.883x-95.96 (R²=0.965, 50≤x≤450 µg). Quantification of tramadol in tissues was compared, at 0.5 or 1 h after injection. Quantifying the trend of tramadol in different tissue was heterogeneous over time. Compared with its quantification at 0.5 h, tramadol in liver and gut tissues, was observably increased as much as 56.4 and 742.2 times in the following 0.5 h, respectively. In eyes, muscle and gill tissues, tramadol was maintained at relatively stable low levels, with only 1.45-fold increase, and 0.88 or 0.92-fold decrease comparing between 0.5 h and 1 h, respectively. However, brain and heart tissue experienced a significant decrease, with 0.096 and 0.01 fold decrease at 1 h compared with 0.5 h (Fig. 3).

The tramadol content in the brain after IM administration was presented in Fig. 4. The standard calibration curve in Fig. 4A (indicated by arrow; y=0.3669x-0.0184; R²=0.999, linear range from 0 to 10 µg), demonstrated a highly linear relationship between the concentration and its peak area in the GC spectra. A 3-exponential model best described the brain concentrations of tramadol for the 10 and 25 µg treatments (n=7). With different doses of IM tramadol administration and detection time, the concentrations in the brain were similar for injection of 10 (Fig. 4A) and 25 µg (Fig. 4B). The curve indicated an initial rapid phase, corresponding to the detection of the tramadol within 1 min, and reached peak value at 5 min (T_max), after a single dose of 25 and 10 µg IM drug injection. The peak contents of tramadol were 1.76 (C_max) and 0.68 µg (C_max) for the 25 and 10 µg injections, respectively, and the ratio between two values was quite similar to that of initial dosage. Despite similar T_max, faster drug clearance was observed with the low-dose group, with the content dropped to initial value (1.11 µg) by 20 min and was detectable up to 3 h in the low-dose group, however, it took 80 min to fall to initial value (1.76 µg) in high-dose groups, and tramadol was detectable up to 4 h after the treatment.

Primary desmethyltramadol reference substances produced using the previously reported method (25), and were analyzed by GC-MS. Compared with the aforementioned strategies to detect tramadol alone, the extraction method and analysis conditions were optimized for combined determination of tramadol and its primary metabolites. The total ion chromatogram (Fig. 5A) and related EI mass spectrograms were obtained (Fig. 5B-D). As presented in Fig. 5A, the three compounds (tramadol, M2 and M1) were separately detected at 9.096, 9.417 and 9.573 min. Data searching in the library, their EI mass spectra consistent with those of tramadol, M2 and M1.

Brain extractions from the IM and bath exposure treatment groups at 0.5, 1 and 2 h were analyzed under the same conditions, and magnified chromatograms are presented in Fig. 6. Three novel peaks were present in all the tramadol treatment groups compared with the control group, even with different treatment strategies, at different sampling times, and using different concentrations. The elution time and EI mass spectra with were consistent with those of the tramadol, M2 and M1 reference samples. Furthermore, distinct changes in tramadol, M1 and M2 concentrations were observed following the two different types of administrations. For example, following IM administration, the concentration of tramadol was very high and gradually decreased between 0.5 and 2 h. A higher concentration of M2 was detected at 1 h compared with 0.5 and 2 h IM treatments. However, in the bath exposure treatment groups, the concentration of tramadol was higher at 1 h than at 0.5 and 2 h. The concentration of M2 was variable, with higher values observed at 0.5 and 2 h, and lower at 1 h. Unfortunately, for M1, data analysis was not always possible due to a very low and variable concentration. Thus, two primary metabolites M1 and M2 were detected, however the total amount of M2 was greater than M1, as the main desmethyl metabolite detected in zebrafish brain.