A novel metabolic pathway of morphine: formation of morphine glucosides in cancer patients

Abstract

Aims

To characterize directly the conjugated metabolites of morphine in urine samples of cancer patients.

Methods

Urine samples from the patients were treated by solid-phase extraction method and chromatographed using three high-performance liquid chromatography systems. Conjugated metabolites were directly detected with liquid chromatographic/ion trap mass spectrometric (LC/MSⁿ) technique by selected ion monitoring, full scan MS/MS and MS³ modes.

Results

Six conjugated metabolites including two new metabolites M5 and M6 were found. Morphine-3-glucuronide (M-3-G) and morphine-6-glucuronide (M-6-G) were identified by comparing their l.c. retention times and multistage mass spectra with those of the reference substances. Two novel metabolites, morphine-3-glucoside and morphine-6-glucoside, as well as normorphine glucuronides were identified by comparing their mass fragment patterns and l.c. retention times with those of M-3-G and M-6-G. Hydrolysis of urine samples with β-glucosidase and β-glucuronidase provided further evidence of the metabolites M5 and M6 as morphine glucosides. The excretion amounts of morphine conjugates in urines were in the order of morphine glucuronides, morphine glucosides and normorphine glucuronides.

Conclusions

In the present study, the applications of l.c. separation and multistage mass spectra have permitted the direct identification of conjugated metabolites of morphine. To our knowledge, this is the first report about O-linked glucosides of morphine at 3-aromatic and 6-aliphatic hydroxyl groups.

Introduction

Morphine, the principal alkaloid of opium, is widely used for the control of moderate to severe pain. Administration by mouth is preferred for terminal cancer pain, whereas parenteral routes are generally used for postoperative pain [1, 2]. The metabolism of morphine has been studied in animals and humans [3–8]. In humans, morphine is predominantly metabolized by glucuronidation in the gut and liver to produce morphine-3-glucuronide (M-3-G) and morphine-6-glucuronide (M-6-G). The latter is considered to contribute to the analgesic effect of morphine, especially when repeated doses are given by mouth [9–11]. There has also been recent interest in the effects of M-3-G [12, 13]. Findings in animals indicate that this metabolite may antagonize the analgesic action of morphine and M-6-G [12]. The two conjugates are excreted rapidly into the urine or bile because of their higher water solubility.

In recent years, with the advent of various soft ionization techniques, especially electrospray ionization (e.s.i.), it has become possible to identify and quantify glucuronides directly with liquid chromatography/mass spectrometry (LC/MS). Morphine and its glucuronides in plasma and urine were simultaneously determined by l.c./quadrupole MS methods [14, 15]. In our laboratory, when we have attempted to determine M-3-G and M-6-G in urine samples from cancer patients by l.c./ion trap MS (LC/MSⁿ) method, new conjugated metabolites of morphine were observed. In the present study, the structures of these conjugates were elucidated by the liquid chromatographic/electrospray multistage mass spectrometric method and enzymatic hydrolysis.

Materials and methods

Chemicals

Morphine hydrochloride was purchased from the National Institute for Control of Pharmaceutical and Biological Products (Beijing, China). M-3-G, M-6-G, β-d-glucosidase (EC 3.2.1.21) and β-D-glucuronidase (EC 3.2.1.31) were purchased from Sigma (St Louis, MO, USA). Methanol and acetonitrile were of high-performance liquid chromatography (HPLC) grade, and other chemicals used were of analytical grade. Distilled water, prepared from demineralized water, was used throughout the study.

Patients and urine collection

Five patients (two females and three males) with terminal cancer of stomach and severe pain gave their informed consent to participation in the study, which was approved by the Ethics Committee of the Second Affiliated Hospital of China Medical University. The mean age of the patients was 67 ± 5 years and the mean weight was 60.6 ± 8.5 kg. Only morphine was administered for pain relief at least 3 days before and during the study period.

The patients were treated with morphine at an oral dose of 10 mg every 4 h for 3 days in this study. Urine was collected for 24 h over wet ice into tubes after the first oral dose on day 3 of the study. Urine samples were stored frozen at −20 °C until analysis.

Extraction of conjugated metabolites of morphine from patient urine

A 1.0-ml portion of urine collected from each patient was applied to Supelclean LC-18 solid-phase extraction column (2 ml; Supelco, Bellefonte, PA, USA) preconditioned with 2-ml aliquots of methanol and water. After loading the urine sample, the column was washed with 1 ml of water. Metabolites were eluted with 1 ml of methanol. The eluate was evaporated to dryness at 40 °C under a gentle stream of nitrogen, and the residue was reconstituted by addition of 200 µl of the mobile phase. A 20-µl aliquot of the solution was injected onto the LC/MSⁿ system.

Enzymatic hydrolysis experiment

Each urine sample obtained from five cancer patients receiving morphine was divided into three 1.0-ml portions. The first portion was incubated for 16 h at 37 °C with β-d-glucosidase (100 U ml⁻¹) after adjusting the pH to 5.0 using 40 mm NH₄H₂PO₄ buffer (pH 3.0). The second portion was incubated for 16 h at 37 °C with β-d-glucuronidase (1000 U ml⁻¹) after adjusting the pH to 5.0. The third portion served as control and was incubated under identical conditions but without the added enzymes for 16 h. Incubations were stopped by using solid-phase extraction described above and analysed by LC/MSⁿ system.

LC/MSⁿ analyses

A Finnigan LCQ ion trap mass spectrometer (San Jose, CA, USA) equipped with an electrospray ionization (ESI) source system, a Shimadzu LC-10AD pump (Kyoto, Japan) and a data system (Xcalibur 1.2) were employed. The interface was adjusted to the following conditions: ion mode, positive; spray voltage, 4.5 kV; capillary temperature, 180 °C; sheath gas (nitrogen), 0.75 l min⁻¹; auxiliary gas (nitrogen), 0.15 l min⁻¹. The full-scan mass spectrum to obtain the protonated molecules [M + H]⁺ of each metabolite was collected in the mass range from m/z 150 to 700. Furthermore, MS/MS and MS³ spectra were obtained for selected precursor ions through incidental collision with neutral gas (helium) molecules in the ion trap, which could provide characteristic fragment ions of each metabolite. The relative collision energy was set at 40%.

The l.c. conditions were as follows (if not specially stated). The column was Nucleosil C₁₈ column (250 × 4.0 mm i.d., 7 µm), supplied by Knauer (Berlin, Germany). The column temperature was maintained at 25 °C. The mobile phase consisted of methanol/water/formic acid, 20/80/1 v/v/v, at a flow rate of 0.5 ml min⁻¹.

Results

Metabolite profiles in urine

The collected urine samples were extracted and analysed as described in Materials and Methods. The total ion current (t.i.c.) chromatogram of a patient urine is displayed in Figure 1A by full scan mode in the mass range from m/z 150 to 700. The parent drug, morphine, could generate a strong protonated molecule [M + H]⁺ at m/z 286 under positive e.s.i. conditions. The selected ion monitoring (s.i.m.) chromatogram by monitoring the [M + H]⁺ ion at m/z 286 in the t.i.c. chromatogram showed three peaks (Figure 1A). Only the peak with the retention time of 9.56 min had the similar l.c. retention time and multistage mass spectra to those of an authentic morphine standard. The full scan MS/MS spectrum of the peak provided a number of characteristic fragment ions at m/z 268 (–H₂O), 229, 211 and 201 (Figure 1B). The parent drug was found in urine samples of all patients. The other two peaks in the s.i.m. chromatogram also had the identical fragment ions to those of morphine, resulting from degradation of morphine conjugates to aglycone in the LC/MS interface.

Figure 1.

Liquid chromatographic/ion trap mass spectrometric (LC/MSⁿ) analysis of morphine and its conjugated metabolites in urine after oral administration of 60 mg day⁻¹ morphine to a cancer patient. (a) Total ion current (t.i.c.) chromatogram and selected ion monitoring (s.i.m.) chromatogram of morphine. (b) Full scan MS/MS spectrum of m/z 286 corresponding to the peak at 9.56 min.

Two major metabolite peaks with the l.c. retention times of 6.89 and 8.48 min were observed on the t.i.c. chromatogram of the urine extract from patients. The two peaks had the identical full scan mass spectra, in which the base peak ions were m/z 462, but the ions at m/z 448 with about 20% of relative abundance were also observed (see Figure 2). The results indicated that each of the two metabolite peaks in the t.i.c. chromatogram corresponded to at least two metabolites. To identify further the structures of these metabolites with the [M + H]⁺ ions at m/z 462 or 448, full scan MS/MS and MS³ modes were applied to obtain the characteristic fragments ions.

Figure 2.

Full scan MS spectra corresponding to the peaks at 6.89 min (a) and 8.48 min (b) in total ion current (t.i.c.) chromatogram

Identification of metabolites with [M + H]⁺ ions at m/z 462

By monitoring the [M + H]⁺ ion at m/z 462, s.i.m. chromatogram revealed two metabolite peaks in the urine extract, which were designated as M1 and M2 (Figure 3A). Compared with the l.c. retention time and multistage mass spectra of the reference substances, M1 and M2 were identified as M-3-G and M-6-G, respectively. Their MS/MS yielded relatively simple spectra with an ion at m/z 286 (−176 Da) deriving from the neutral loss of glucuronic acid moiety. Further full scan MS³ spectra of m/z 462→286 produced fragment ions at m/z 268, 229, 211 and 201, which were identical to those produced from the MS/MS spectrum of morphine. The chromatographic behaviours showed that morphine glucuronide conjugated at the 3-OH group is more polar than that conjugated at the 6-OH group.

Figure 3.

Selected ion monitoring (s.i.m.) result for morphine glucuronides (protonated molecule of m/z 462) (a) and full scan MS/MS spectra of m/z 462 corresponding to the peaks at 6.85 min (b) and 8.55 min (c)

Identification of metabolites with [M + H]⁺ ions at m/z 448

The [M + H]⁺ ions at m/z 448, 162 Da higher than that of the parent drug and 14 Da lower than that of morphine glucuronides, suggested that the corresponding metabolites could be normorphine glucuronides or morphine glucosides. The s.i.m. chromatogram (m/z 448, Figure 4A) showed two peaks in the urine extract, designated as peak I and peak II, respectively. The two peaks had similar multistage MS spectra. It was observed that the first half and second half of each peak corresponded to the different MS/MS spectra, indicating that each contained at least two metabolites. The MS/MS spectrum of peak I eluting from 6.5 to 6.8 min yielded major fragment ions at m/z 286, 272 and 254 (Figure 4B), whereas MS/MS spectrum corresponding to the same peak eluting from 7.0 to 7.5 min formed a major fragment ion at m/z 286, and only a small amount of fragment ion at m/z 272 was observed (Figure 4C). Similar to peak I, peak II might also contain two metabolites, one producing the major fragment ions at m/z 272 and 254, another giving a major fragment ion at m/z 286 (Figure 4D,E). The metabolites that could produce the ions at m/z 272 and 254 were designated as M3 and M4, respectively, whereas the metabolites with the characteristic fragment ions of m/z 286 were designated as M5 and M6.

Figure 4.

Selected ion monitoring (s.i.m.) result for morphine glucosides and normorphine glucuronides (protonated molecule at m/z 448) (a) and full scan MS/MS spectra of peak I eluting from 6.5 to 6.8 min (b) (M3), of peak I eluting from 7.0 to 7.5 min (c) (M5), of peak II eluting from 8.1 to 8.4 min (d) (M4), and of peak II eluting from 8.6 to 9.0 min (e) (M6)

The fragment ions at m/z 272 and 254 could be explained by the neutral loss of glucuronic acid moiety (−176 Da) from the precursor ion and the loss of water (−18 Da) from the aglycone, respectively. Furthermore, full scan MS³ spectra of 448→272 both yielded a major fragment ion at m/z 254 as well as the characteristic fragment ions of morphine at m/z 229, 211 and 201 (Figure 5). The precursor ions (m/z 448) and major fragment ions (m/z 272, aglycone) of M3 and M4 were 14 Da lower than those for morphine glucuronides, indicating that their structures were normorphine glucuronides. Similar to morphine, normorphine also has two potential sites for glucuronidation. According to the elution order of M-3-G and M-6-G on the C₁₈ column, the structures of M3 and M4 were proposed as normorphine-3-G and normorphine-6-G, respectively.

Figure 5.

Full scan MS³ spectrum of m/z 448→272 from the peaks of normorphine glucuronides

The fragment ion at m/z 286 from peaks I and II was 162 lower than that of the precursor ion at m/z 448, indicating the neutral loss of glucose moiety. Furthermore, the MS³ spectrum of 448→286 produced the identical fragment ions to those produced from MS/MS spectrum of morphine. The MS data suggested M5 and M6 corresponded to morphine glucosides, which are novel metabolites of morphine found in cancer patients. As the reference substances of M5 and M6 were not available, the structures of M5 and M6 were proposed as morphine-3-glucoside and morphine-6-glucoside, respectively, according to the elution order of M-3-G and M-6-G on the C₁₈ column.

Six conjugated metabolites were found in urine samples of cancer patients by the l.c. separation and multistage MS analysis. The selected reaction monitoring (s.r.m.) chromatograms of these conjugated metabolites are displayed in Figure 6 using the transitions m/z 462→286, m/z 448→286 and m/z 448→272. Each scan channel corresponded to two metabolites, which were 3-O- and 6-O-conjugated isomers. Metabolites M1, M3 and M5 have similar l.c. retention times at about 6.8 min, but these metabolites could be separated according to [M + H]⁺ ions or characteristic fragment ions. Metabolites M2, M4 and M6 possessed similar retention times at about 8.5 min under the l.c. conditions. To confirm further the structures of these conjugated metabolites, two experiments were carried out. First we investigated the cleavage of these putative morphine conjugates by enzyme treatment, and second, we tried to resolve these metabolites using different chromatographic systems.

Figure 6.

Selected reaction monitoring (s.r.m.) results of morphine glucuronide, morphine glucosides and normorphine glucuronides using the transitions of m/z 462→286, m/z 448→286 and m/z 448→272, respectively. Urine sample of a cancer patient after solid-phase extraction was chromatographed on a Nucleosil C₁₈ column.

Enzymatic hydrolysis experiment

In order to investigate further the structures of metabolites M3–M6, urine samples containing these metabolites were subjected to enzymatic hydrolysis. The results are expressed as the percentage of metabolite that was hydrolysed with β-glucuronidase or β-glucosidase to the control (without enzymes). After urine samples were incubated with β-glucuronidase, the amounts of M1–M4 in urines were reduced by more than 93%, while those of M5 and M6 were reduced by about 45%. Treatment with β-glucosidase led to more than 89% reduction of M5 and M6, while the amounts of M1–M4 did not change. These results further supported that M3 and M4 were metabolites conjugated with glucuronic acid, whereas M5 and M6 were glucose conjugates. Proposed structural formulae of the conjugated metabolites of morphine in patient urines are summarized in Figure 7.

Figure 7.

Glucuronic and glucosidic mono-conjugated metabolites of morphine in urine of cancer patients

Chromatographic resolution of morphine conjugates

Several chromatographic systems were applied to achieve the chromatographic resolution of six conjugated metabolites of morphine. S.r.m. mode was applied to detect metabolites M1/M2, M3/M4 and M5/M6 using the transitions of m/z 462→286, m/z 448→272 and m/z 448→286, respectively. On C₁₈ columns, the isomers of morphine conjugates could be separated chromatographically from each other, but the different types of conjugates, e.g. morphine glucuronide, normorphine glucuronide and morphine glucoside, could not be separated. In contrast, on a Spherisorb NH₂ column (250 × 4.6 mm i.d., 10 µm; Elite Inc., Dalian, China), glucuronide and glucoside conjugates of morphine could be separated using methanol/water/formic acid (5/95/0.5 v/v/v) as the mobile phase at a flow rate of 0.5 ml min⁻¹, but their isomers could not. So in the study, a tandem combination of Nucleosil C₁₈ column and Spherisorb NH₂ column was used to resolve six conjugated metabolites of morphine in urines. The mobile phase consisted of methanol/water/formic acid, 10/90/0.1, v/v/v, at a flow rate of 0.4 ml min⁻¹. As a result, the tandem column chromatographic method allowed us to resolve six conjugates of morphine. Their corresponding t.i.c. and s.r.m. chromatograms as well as s.i.m. chromatogram of the parent drug are displayed in Figure 8. Six conjugated metabolites and morphine could be detected in urine samples of all cancer patients, except that morphine-6-glucoside was not found in a cancer patient. It was found that the urine concentrations of higher polar M-3-G and morphine-3-glucoside were higher than those of their 6-isomers, but the opposite tendency was observed for normorphine glucuronides. Among six conjugated metabolites in urine samples, M-3-G had the highest excretion amount. Peak area ratios for conjugated metabolites to M-3-G in urine are listed in Table 1.

Figure 8.

Liquid chromatographic/ion trap mass spectrometric (LC/MSⁿ) analysis of morphine and its conjugated metabolites in urine after oral administration of 60 mg day⁻¹ morphine to a cancer patient. The urine sample was treated by solid-phase extraction, and chromatographed on tandem Nucleosil C₁₈–Spherisorb NH₂ columns using methanol/water/formic acid (10/90/0.1, v/v/v) as the mobile phase, at a flow-rate of 0.4 ml min⁻¹.

Table 1.

Chromatographic peak area (full scan analysis) ratios for conjugated metabolites to M-3-G in urine samples collected from five cancer patients receiving morphine (60 mg day⁻¹)

Patients	M1	M2	M3	M4	M5	M6	Morphine
A	100	14.1	0.12	0.29	1.65	0.15	3.93
B	100	30.3	0.15	0.80	2.03	0.33	13.9
C	100	26.3	0.12	0.70	4.94	0.16	14.9
D	100	22.6	0.071	0.18	2.30	n.d.	14.7
E	100	16.0	0.10	0.22	6.45	0.97	23.7
Mean	100	21.9	0.11	0.44	3.47	0.40	14.2
SD		6.81	0.03	0.29	2.11	0.39	7.02

Although there might be significant differences in the MS responses of these conjugated metabolites, the data listed in Table 1 could provide a sequence in the excretion amount of the conjugates in urines. Accordingly, the excretion amounts of M-3-G and M-6-G in each patient urine both exceeded the level of the parent drug. The excretion of morphine glucosides relative to glucuronides exhibited significant interindividual variation. The amount of morphine-3-glucoside was 15 times higher than that of its isomer. The peak area of morphine glucosides exceeded that of normorphine glucuronides in each patient, and their ratios also exhibited significant interindividual variation.

Discussion

It is well recognized that conjugations are important metabolic pathways of many xenobiotic and endogenous compounds. There is a growing interest in identifying such metabolites and in investigating their pharmacokinetic, pharmacological and toxicological properties. In comparison with glucuronidation, glucosidation represents a relatively minor conjugation process in humans. There is only very limited information in the literature on this topic. The reported glucoside conjugates were focused on the N-linked glucosides formed through primary and secondary amines of the drugs, e.g. 5-aminosalicylic acid [16], amobarbital [17] and phenobarbital [18]. O-linked glucoside metabolites of drugs were rarely found in humans. The glucoside of mycophenolic acid was the first O-linked glucoside of a drug identified in humans, which was formed through carboxyl group [19].

In the present study, the applications of HPLC separation and multistage MS have permitted the direct identification of conjugated metabolites of morphine. Using these methods, it was shown that four glucuronide conjugates (M-3-G, M-6-G and normorphine glucuronides) and two glucoside conjugates (morphine-3-glucoside and morphine-6-glucoside) were excreted into urine after the oral administration of morphine to cancer patients (Figure 7). Among these metabolites, the conjugates with glucuronic acid have been reported in the literature [3–8], but to our knowledge there was no report about O-linked glucosides of morphine at 3-aromatic and 6-aliphatic hydroxyl groups. Similar to M-3-G and M-6-G, the amount of morphine-3-glucoside excreted into urine after administration of morphine is greater than that of morphine-6-glucoside. These data imply that the aromatic hydroxyl group of morphine shows stronger metabolic reactivity than the aliphatic hydroxyl group, but the opposite is true for normorphine. The clinical significance of the glucoside conjugates and the relationship between the formation of the conjugates and cancer remain to be investigated.