Quantitative Analysis of Tiamenidine in Human Plasma by Gas Chromatography Mass Spectrometry of a Dibenzyl Derivative T. A. Bryce and J. L. Burrows Hoechst Pharmaceutical Research Laboratories, Walton Manor, Walton, Milton Keynes, Buckinghamshire MK7 7AJ,UK
A gas chromatography mass spectrometry method has been developed and evaluated for the quantitative analysis of tiamenidine in plasma. Tiamenidine and internal standard are extracted from hasilied plasma, converted to dibenzyl derivatives by reaction with benzyl bromide and potassium t-butoxide in the presence of 18-crown-6 ether prior to analysis by selected ion monitoring. The method can be used over the range 0.2-10 ng ml-' with a coefficient of variation of better than 20% at 1 ng mi-'.
INTRODUCTION Tiamenidine hydrochloride (2-[(2-chloro-4-methyl-3thienyl)amino]-2-imidazolinehydrochlorine, also HOE 440) (1) is a potent antihypertensive in man.' As the drug is normally administered in low doses (1-2 mg) a highly sensitive analytical method is needed to investigate its kinetics and metabolism. H
1
also arise in the preparation and handling of the micropacked glass columns described in this method. As it has been shown the N-alkyl derivatives of clonidine are stable, it was decided to investigate the preparation of N-alkyl derivatives of tiamenidine. Consequently, a GCMS assay of tiamenidine has been developed based on the conversion of tiamenidine using a crown ether catalyst to a stable di-N-benzyl derivative which can be analysed using conventional packed columns. No special deactivation of the GCMS interface is required to obtain sensitivities of 0.2 ng ml-'
EXPERIMENTAL
2
Clonidine (2-(2,6-dichlorophenylamino)-2-imidazoline) which is structurally similar to tiamenidine, has been analysed after therapeutic doses by two methods. In the first method,* clonidine is methylated by 'on-column' reaction using the reagent trimethylanilinium hydroxide and quantified by GCMS, while the second method3 quantifies clonidine as its N-pentafluorobenzyl derivative by G C with electron capture detection. A recent r e p ~ r t as , ~well as pointing out the unsuitability of the 'on-column' methylation technique for tiamenidine, has described a method for the quantitative analysis of tiamenidine by GCMS of its bisheptafluorobutyryl derivative. While the sensitivity of this method (lower limit of detection 0.2 ng ml-') is sufficient to measure plasma levels of tiamenidine after a single oral dose of 1 mg, there are a number of difficulties which could arise in the application of the method to a large number of clinical samples. These difficulties are due mainly to the instability of the bisheptafluorobutyryl derivative, which must be stored for analysis in a solution containing heptafluorobutyric anhydride. Repeated injection of the anhydride, which probably also contains small amounts of the corresponding acid, severely reduces the life of the G C column. In addition, the GCMS interface must be deactivated with large amounts of hexamethyldisilazane to obtain optimum performance. Other difficulties could
Materials and methods All solvents and reagents were of analytical grade and were used without further purification. Benzyl bromide and 18-crown-6 ether (1,4,7,10,13,16-hexaoxacyclooctadecane) were purchased from Aldrich Chemical Co. Standard solutions of tiamenidine hydrochloride (25 p g ml-') and internal standard, desmFthy1tiamenidine hydrochloride (2) (100 ,xg ml- ) in 0.01 M HCl were prepared by dissolving the solid material in 1 M HCl and diluting with distilled water to the appropriate volumes. Portions of these solutions were diluted with 0.01 M HCl to concentrations of 0.2 p g ml-' and 1 p g ml-' of tiamenidine hydrochloride and desmethyltiamenidine hydrochloride, respectively. Test tubes were silanized with a 5 % (v/v) solution of dichlorodimethylsilane in hexane, rinsed in methanol and dried before use. In addition to reducing any adsorption losses of tiamenidine, this procedure also assisted the separation of organic and aqueous phases during extraction. Extractions were carried out in 10 or 15 ml screwcapped test tubes using a rotary inversion mixer at a fixed speed of 20 rev min-' for a period of 10 min. After extraction, the phases were separated by centrifuging the test tubes at a speed of 2500 rev min-' (c. 2000 g).
CCC-0306-042X/79/0006-0027$02.00 @ Heyden & Son Ltd, 1979
BIOMEDICAL MASS SPECTROMETRY, VOL. 6, NO. 1, 1979 27
T. A. BRYCE AND J. L. BURROWS
Extraction and derivatization To plasma (4 ml) in a screw-capped test-tube is added 100 ng of internal standard (100 p1 of the 1pg ml-' standard solution), dichloromethane (5 ml) and 1M NaOH (1ml). After extraction and centrifugation, the upper aqueous phase is aspirated and discarded. The dichloromethane which remains is transferred to a clean test tube and evaporated to dryness at 40°C under a gentle stream of nitrogen. The residue is taken up in hexane (1 ml). Solid potassium f-butoxide (5-10 mg), 18-crown-6 ether (0.5mg in 1 0 0 ~ 1of hexane) and benzyl bromide (10 PI) are added. The contents of the tube are mixed with the aid of a vortex mixer, and reacted at 60 "C for 30 min in a water bath. The cooled reaction mixture is diluted with hexane (5 ml) and the benzyl derivatives are extracted into 1 M HCl (1 ml). The organic phase is aspirated and discarded. The aqueous phase is washed once with hexane (2-3 ml), and the benzyl derivatives are reextracted into cyclohexane (5 ml), after the aqueous phase has been made alkaline by adding 1 M NaOH (1.5 ml). The cyclohexane is then transferred to a tapered test tube and evaporated to dryness under a gentle stream of nitrogen at 40°C. The residue is dissolved in butyl acetate (10 p l ) and aliquots (5-8 p I ) are analysed by GCMS.
Gas chromatography mass spectrometry A Pye 104 gas chromatograph coupled via a singlestage, all-glass jet separator to an AEI MS 30 double beam mass spectrometer was used for the analyses. The mass spectrometer was equipped with a multipeak monitoring accessory. Gas chromatography was performed on a 1.5 m X 2 mm ID glass column packed with 3% OV-1 on Chromosorb W-HP (100/120 mesh). The column temperature was 275 "C, and the injection port and interface temperatures were c. 300 "C and c. 240 "C, respectively. The carrier gas was helium flowing at 30 ml min-'. Retention times of the dibenzyl derivatives of tiamenidine and desmethyltiamenidine were 2.8 and 2.6 min respectively, under these conditions. The mass spectrometer source temperature was 200 OC, the ionizing current was 300 p A, the electron beam energy was 70 eV, and the resolving power was 1000. The multipeak monitoring accessory was used to record the ion currents at m / z 360 and m / z 346, i.e. the [M - Cl]+ ions and dibenzyltiamenidine and dibenzyldesmethyltiamenidine, respectively. Selected ion chromatograms obtained from blank plasma and from plasma to which had been added tiamenidine hydrochloride (1 ng ml-') are shown in Fig. 1.
1
1
-0
-0
6
6
Time (rnin)
Figure 1. Examples of selected ion chrornatograrns from human plasma: (a) plasma to which had been added desmethyltiamenidine hydrochloride (25 ng rn1-l); (b) plasma to which had been added tiarnenidine hydrochloride (1 ng rn1-l) and desrnethyltiarnenidine hydrochloride (25 ng ml-'). The peaks due to the dibenzyl derivatives of desrnethyltiamenidine and tiamenidine are indicated by 1 and 2, respectively. Note that the recorder pens are off-set.
tiamenidine. This response factor was determined experimentally for each batch of samples by analysing, in parallel with the unknown samples, blank plasma (4 ml) to which had been added 20 ng of tiamenidine hydrochloride (100 p l of the 0.2 p g ml-' standard solution) as well as internal standard.
Extraction from plasma The extraction. efficiency of tiamenidine from basified plasma was investigated using [14C]tiamenidine as follows. A solution of ['4C]tiamenidine (0.23 pg, 12 nCi) in 0.01 M HCl (100 p l ) was mixed with plasma (2 ml) and various organic solvents (5 ml) and 1M NaOH (1 ml), or buffer (PH 10, 1 ml), were added. The plasma was extracted for 20 min and the phases separated by centrifugation, as described previously. A portion (3 ml) of the organic phase was transferred to a counting vial. An ether solution of hydrogen chloride was added to the halogenated solvents which were then evaporated to dryness under nitrogen and the residues were taken up in methanol (1 ml). Scintillator was added to all the samples and the amount of radioactivity recovered was determined by liquid scintillation counting.
Quantification of tiamenidine hydrochloride
RESULTS AND DISCUSSION
The concentration of tiamenidine hydrochloride in a plasma sample was determined from the peak height ratio of m / z 360 to m / z 346, and from the response factor of dibenzyltiamenidine to dibenzyldesmethyl-
Extraction from plasma
28 BIOMEDICAL MASS SPECTROMETRY, VOL. 6, NO. 1, 1979
The results of the recovery experiments (Table 1)show that recoveries with hydrocarbon solvents are generally @ Heyden & Son Ltd, 1979
QUANTITATIVE ANALYSIS OF TIAMENIDINE ~
Table 1. The extraction of ["C] tiamenidme from basled plasma into various organic solvents.
Table2 Retention indices of the four benzyl derivatives of tiamenidine Derivative
Solvent
Activity extracted into organic phase (%) (duplicatedeterminations) pH lobuffera 1 M NaOH
Pentane Hexane Cyclohexane Toluene Chloroform Dichloromethane 1.2-Dichloroethane Diethyl ether Ethyl acetate
0.1,O.l 0.2,O.l 0.4,0.4 17,17 83,80 74,69 52,48 29,24 42,53
2.0,2.0 2.7,2.9 3.7,3.7 75,76 91,95 98,99 93,94 90,91 98,96
"0.1 M glycine+NaOH+NaCI.
poor. Good values are obtained using oxygenated and halogenated solvents, especially when the aqueous medium is made strongly alkaline with 1 M NaOH. The best values resulted from using dichloromethane, or 1,2-dichloroethane, and the former was chosen for the extraction of tiamenidine because of its lower boiling point, and consequent ease of removal by evaporation. Similarly, it was shown that ["C] tiamenidine is most efficiently extracted from plasma which has been basified with 1 M NaOH when extraction times of at least 10 min are used.
Derivatization As it has been shown previously4 that fluoroacyl derivatives of tiamenidine are unstable, alkylation of tiamenidine was investigated. A stable pentafluorobenzyl derivative of tiamenidine was prepared using the conditions described for ~ l o n i d i n eHowever, .~ the response of the tiamenidine derivative to electron capture detection was not as great as that of the clonidine derivative, and this approach was abandoned as the predicted sensitivity was unlikely to be sufficient to allow monitoring of tiamenidine plasma levels. In addition, it was thought desirable to prepare a dialkyl derivative, since the chromatographic properties were expected to be superior to those of a monoalkyl derivative. It was found that reacting tiamenidine (free base) in acetone with excess benzyl bromide at 50 "C for 90 min in the presence of sodium carbonate, potassium carbonate or potassium t-butoxide resulted in the formation of a mixture of two monobenzyl and, because of tautomerization, two dibenzyl derivatives, the relative amounts of each compound being dependent on the base used. The four alkyl derivatives could be readily separated by GC, and their retention indices are shown in Table 2. The molecular weight of each compound was confirmed by GCMS. It was not possible to determine the substitution patterns from the mass spectra. When the same series of reactions was carried out using benzene as solvent and after adding 18-crown-6 ether as catalyst, it was found that the dibenzyl derivative with the gxeater retention time was formed almost exclusively (>goo/,) when potassium t-butoxide was the base; whereas the crown ether had little effect on the @ Heyden & Son Ltd, 1979
Retention index'
2290 2500 2780 2970
Monobenzyl Monobenzyl Dibenzyl Dibenzyl
a The retention indices were measured on a 1.8 m x 4 mm i.d. glass column packed with 3% OV-l on Chromosorb W-HP (100/120 mesh). The oven temperature was 250 "C and the carrier (nitrogen) flow was 60 ml min-'.
relative amounts of products from the reactions using sodium or potassium carbonate. ' T h e reaction of tiamenidine with benzyl bromide in the presence of 18-crown-6 ether proceeds equally well in hexane and this solvent was preferred in the final method. Crown ethers have been used previously to prepare phenacyf and pentaflu~robenzyl~ esters for chromatography. Generally the esters are obtained in qualitative yields (>97%) free from by-products. The extension, reported here, of this type of reaction to the formation of N-alkyl derivatives could well have wide application.
Mass spectra The mass spectra of the dibenzyl derivatives of tiamenidine and desmethyltiamenidine formed in the presence of 18-crown-6 ether are shown in Figs. 2 and 3. As the molecular ions are small, the [M-C1]+ ions, which account for approximately 25% of the total ionization in each case, are monitored in the method.
Accuracy and precision The accuracy and precision of the method was determined by analysing plasma to which had been added tiamenidine hydrochloride over the range 0.2 ng m1-I10 ng rn1-l. Each sample was analysed six times and the results obtained are shown in Table 3. Comparison of 1001
I
m/z
Figure 2. Mass spectrum of the dibenzyl derivative of tiamenidine formed in the presence of 18-crown-6 ether.
BIOMEDICAL MASS SPECTROMETRY, VOL. 6, NO. 1. 1979 29
T. A. BRYCE AND J. L. BURROWS
91
Table3. Analysis of plasma after the addition of various amounts of tiamenidine hydrochloride 3L6
rl
* I
1 " " I " ~ ~ l
200 m /I
I,
,!, , ,
,,
,
300
,,
1 a
Tiamenidine hydrochloride added (ng rn1-l)
Tiamenidine hydrochloride found Mean SD (6 determinations) (ng m1-l)
0 0.21 0.52 1.01 2.04 5.20 10.3
0.06* 0.04 0.25* 0.03 0.55*0.04 1.11*0.19 2.02*0.17 4.98*0.31 10.8 *0.6
*
Recovery of tiamenidine hydrochloridea
1%)
90* 12 94* 7 104*17 96* 8 95+ 6 104* 6
After subtraction of blank value of 0.06 ng ml-'
Figure 3. Mass spectrum of the dibenzyl derivative of desrnethyltiamenidine formed in the presence of 18-crown-6 ether.
the mean determined tiamenidine concentrations with the actual tiamenidine levels shows that the accuracy is within *looh.The precision was found to be better than *20% for levels of 1 ng ml-' and below, and better than *lo% for levels greater than 1 ng ml-'.
Sensitivity When blank plasma samples were analysed, a small peak was always observed [Fig. 2(A)] at the retention time
of dibenzyltiamenidine. It has not been established whether this peak represents a co-extracted compound or is derived from dibenzyltiamenidine washed out from the syringe, column or interface. However, the average amount of this peak was found to be (Table 3) equivalent to 0.06 ng ml-' tiamenidine hydrochloride. Although this limits the sensitivity of the method, satisfactory values of accuracy and precision were obtained for concentrations of 0.2 ng m1-l.
REFERENCES 1. F. Kersting,Arzneim.-forsch.23,1657 (1973);E. Lindnerand J. Kaiser, Arch. Int. Pharmacodyn. Ther. 221, 305 (1974). 2. G. H. Draffan, R. A. Clare, S.Murray, G. D. Bellward, D. S. Davies and C. T. Dollery, in Advance in Mass Spectrometry in Biochemistry and Medicine, Vol. 2, ed. by A. Frigerio, p. 389, Spectrum Publications, New York (1977). 3. P.-0. Edldnd and L. K. Paalzow. Acta Pharmacol. Toxicol. 40-, 145 ( 1977). 4. H.-W. Fehlhaber, K. Metternich, D. Tripier and M. Uihlein, Biomed. Mass Spectrom, 5, 188 (1978).
30 BIOMEDICAL MASS SPECTROMETRY, VOL. 6, NO. 1, 1979
5. H. Stahle and K.-H. Pook, LiebigsAnn. Chem. 751, 159 (1971). 6. H. D. Hurst, M. Milano, E. J. Kikta Jr, S. A. Connelly and E. Grushka, Anal. Chem. 47, 1797 (1975). 7. 8. Davis, Anal. Chem. 49,832 (1977).
Received 26 July 1978 @ Heyden & Son Ltd, 1979
@ Heyden & Son Ltd, 1979
A gas chromatography mass spectrometry method has been developed and evaluated for the quantitative analysis of tiamenidine in plasma. Tiamenidine and internal standard are extracted from hasilied plasma, converted to dibenzyl derivatives by reaction with benzyl bromide and potassium t-butoxide in the presence of 18-crown-6 ether prior to analysis by selected ion monitoring. The method can be used over the range 0.2-10 ng ml-' with a coefficient of variation of better than 20% at 1 ng mi-'.
INTRODUCTION Tiamenidine hydrochloride (2-[(2-chloro-4-methyl-3thienyl)amino]-2-imidazolinehydrochlorine, also HOE 440) (1) is a potent antihypertensive in man.' As the drug is normally administered in low doses (1-2 mg) a highly sensitive analytical method is needed to investigate its kinetics and metabolism. H
1
also arise in the preparation and handling of the micropacked glass columns described in this method. As it has been shown the N-alkyl derivatives of clonidine are stable, it was decided to investigate the preparation of N-alkyl derivatives of tiamenidine. Consequently, a GCMS assay of tiamenidine has been developed based on the conversion of tiamenidine using a crown ether catalyst to a stable di-N-benzyl derivative which can be analysed using conventional packed columns. No special deactivation of the GCMS interface is required to obtain sensitivities of 0.2 ng ml-'
EXPERIMENTAL
2
Clonidine (2-(2,6-dichlorophenylamino)-2-imidazoline) which is structurally similar to tiamenidine, has been analysed after therapeutic doses by two methods. In the first method,* clonidine is methylated by 'on-column' reaction using the reagent trimethylanilinium hydroxide and quantified by GCMS, while the second method3 quantifies clonidine as its N-pentafluorobenzyl derivative by G C with electron capture detection. A recent r e p ~ r t as , ~well as pointing out the unsuitability of the 'on-column' methylation technique for tiamenidine, has described a method for the quantitative analysis of tiamenidine by GCMS of its bisheptafluorobutyryl derivative. While the sensitivity of this method (lower limit of detection 0.2 ng ml-') is sufficient to measure plasma levels of tiamenidine after a single oral dose of 1 mg, there are a number of difficulties which could arise in the application of the method to a large number of clinical samples. These difficulties are due mainly to the instability of the bisheptafluorobutyryl derivative, which must be stored for analysis in a solution containing heptafluorobutyric anhydride. Repeated injection of the anhydride, which probably also contains small amounts of the corresponding acid, severely reduces the life of the G C column. In addition, the GCMS interface must be deactivated with large amounts of hexamethyldisilazane to obtain optimum performance. Other difficulties could
Materials and methods All solvents and reagents were of analytical grade and were used without further purification. Benzyl bromide and 18-crown-6 ether (1,4,7,10,13,16-hexaoxacyclooctadecane) were purchased from Aldrich Chemical Co. Standard solutions of tiamenidine hydrochloride (25 p g ml-') and internal standard, desmFthy1tiamenidine hydrochloride (2) (100 ,xg ml- ) in 0.01 M HCl were prepared by dissolving the solid material in 1 M HCl and diluting with distilled water to the appropriate volumes. Portions of these solutions were diluted with 0.01 M HCl to concentrations of 0.2 p g ml-' and 1 p g ml-' of tiamenidine hydrochloride and desmethyltiamenidine hydrochloride, respectively. Test tubes were silanized with a 5 % (v/v) solution of dichlorodimethylsilane in hexane, rinsed in methanol and dried before use. In addition to reducing any adsorption losses of tiamenidine, this procedure also assisted the separation of organic and aqueous phases during extraction. Extractions were carried out in 10 or 15 ml screwcapped test tubes using a rotary inversion mixer at a fixed speed of 20 rev min-' for a period of 10 min. After extraction, the phases were separated by centrifuging the test tubes at a speed of 2500 rev min-' (c. 2000 g).
CCC-0306-042X/79/0006-0027$02.00 @ Heyden & Son Ltd, 1979
BIOMEDICAL MASS SPECTROMETRY, VOL. 6, NO. 1, 1979 27
T. A. BRYCE AND J. L. BURROWS
Extraction and derivatization To plasma (4 ml) in a screw-capped test-tube is added 100 ng of internal standard (100 p1 of the 1pg ml-' standard solution), dichloromethane (5 ml) and 1M NaOH (1ml). After extraction and centrifugation, the upper aqueous phase is aspirated and discarded. The dichloromethane which remains is transferred to a clean test tube and evaporated to dryness at 40°C under a gentle stream of nitrogen. The residue is taken up in hexane (1 ml). Solid potassium f-butoxide (5-10 mg), 18-crown-6 ether (0.5mg in 1 0 0 ~ 1of hexane) and benzyl bromide (10 PI) are added. The contents of the tube are mixed with the aid of a vortex mixer, and reacted at 60 "C for 30 min in a water bath. The cooled reaction mixture is diluted with hexane (5 ml) and the benzyl derivatives are extracted into 1 M HCl (1 ml). The organic phase is aspirated and discarded. The aqueous phase is washed once with hexane (2-3 ml), and the benzyl derivatives are reextracted into cyclohexane (5 ml), after the aqueous phase has been made alkaline by adding 1 M NaOH (1.5 ml). The cyclohexane is then transferred to a tapered test tube and evaporated to dryness under a gentle stream of nitrogen at 40°C. The residue is dissolved in butyl acetate (10 p l ) and aliquots (5-8 p I ) are analysed by GCMS.
Gas chromatography mass spectrometry A Pye 104 gas chromatograph coupled via a singlestage, all-glass jet separator to an AEI MS 30 double beam mass spectrometer was used for the analyses. The mass spectrometer was equipped with a multipeak monitoring accessory. Gas chromatography was performed on a 1.5 m X 2 mm ID glass column packed with 3% OV-1 on Chromosorb W-HP (100/120 mesh). The column temperature was 275 "C, and the injection port and interface temperatures were c. 300 "C and c. 240 "C, respectively. The carrier gas was helium flowing at 30 ml min-'. Retention times of the dibenzyl derivatives of tiamenidine and desmethyltiamenidine were 2.8 and 2.6 min respectively, under these conditions. The mass spectrometer source temperature was 200 OC, the ionizing current was 300 p A, the electron beam energy was 70 eV, and the resolving power was 1000. The multipeak monitoring accessory was used to record the ion currents at m / z 360 and m / z 346, i.e. the [M - Cl]+ ions and dibenzyltiamenidine and dibenzyldesmethyltiamenidine, respectively. Selected ion chromatograms obtained from blank plasma and from plasma to which had been added tiamenidine hydrochloride (1 ng ml-') are shown in Fig. 1.
1
1
-0
-0
6
6
Time (rnin)
Figure 1. Examples of selected ion chrornatograrns from human plasma: (a) plasma to which had been added desmethyltiamenidine hydrochloride (25 ng rn1-l); (b) plasma to which had been added tiarnenidine hydrochloride (1 ng rn1-l) and desrnethyltiarnenidine hydrochloride (25 ng ml-'). The peaks due to the dibenzyl derivatives of desrnethyltiamenidine and tiamenidine are indicated by 1 and 2, respectively. Note that the recorder pens are off-set.
tiamenidine. This response factor was determined experimentally for each batch of samples by analysing, in parallel with the unknown samples, blank plasma (4 ml) to which had been added 20 ng of tiamenidine hydrochloride (100 p l of the 0.2 p g ml-' standard solution) as well as internal standard.
Extraction from plasma The extraction. efficiency of tiamenidine from basified plasma was investigated using [14C]tiamenidine as follows. A solution of ['4C]tiamenidine (0.23 pg, 12 nCi) in 0.01 M HCl (100 p l ) was mixed with plasma (2 ml) and various organic solvents (5 ml) and 1M NaOH (1 ml), or buffer (PH 10, 1 ml), were added. The plasma was extracted for 20 min and the phases separated by centrifugation, as described previously. A portion (3 ml) of the organic phase was transferred to a counting vial. An ether solution of hydrogen chloride was added to the halogenated solvents which were then evaporated to dryness under nitrogen and the residues were taken up in methanol (1 ml). Scintillator was added to all the samples and the amount of radioactivity recovered was determined by liquid scintillation counting.
Quantification of tiamenidine hydrochloride
RESULTS AND DISCUSSION
The concentration of tiamenidine hydrochloride in a plasma sample was determined from the peak height ratio of m / z 360 to m / z 346, and from the response factor of dibenzyltiamenidine to dibenzyldesmethyl-
Extraction from plasma
28 BIOMEDICAL MASS SPECTROMETRY, VOL. 6, NO. 1, 1979
The results of the recovery experiments (Table 1)show that recoveries with hydrocarbon solvents are generally @ Heyden & Son Ltd, 1979
QUANTITATIVE ANALYSIS OF TIAMENIDINE ~
Table 1. The extraction of ["C] tiamenidme from basled plasma into various organic solvents.
Table2 Retention indices of the four benzyl derivatives of tiamenidine Derivative
Solvent
Activity extracted into organic phase (%) (duplicatedeterminations) pH lobuffera 1 M NaOH
Pentane Hexane Cyclohexane Toluene Chloroform Dichloromethane 1.2-Dichloroethane Diethyl ether Ethyl acetate
0.1,O.l 0.2,O.l 0.4,0.4 17,17 83,80 74,69 52,48 29,24 42,53
2.0,2.0 2.7,2.9 3.7,3.7 75,76 91,95 98,99 93,94 90,91 98,96
"0.1 M glycine+NaOH+NaCI.
poor. Good values are obtained using oxygenated and halogenated solvents, especially when the aqueous medium is made strongly alkaline with 1 M NaOH. The best values resulted from using dichloromethane, or 1,2-dichloroethane, and the former was chosen for the extraction of tiamenidine because of its lower boiling point, and consequent ease of removal by evaporation. Similarly, it was shown that ["C] tiamenidine is most efficiently extracted from plasma which has been basified with 1 M NaOH when extraction times of at least 10 min are used.
Derivatization As it has been shown previously4 that fluoroacyl derivatives of tiamenidine are unstable, alkylation of tiamenidine was investigated. A stable pentafluorobenzyl derivative of tiamenidine was prepared using the conditions described for ~ l o n i d i n eHowever, .~ the response of the tiamenidine derivative to electron capture detection was not as great as that of the clonidine derivative, and this approach was abandoned as the predicted sensitivity was unlikely to be sufficient to allow monitoring of tiamenidine plasma levels. In addition, it was thought desirable to prepare a dialkyl derivative, since the chromatographic properties were expected to be superior to those of a monoalkyl derivative. It was found that reacting tiamenidine (free base) in acetone with excess benzyl bromide at 50 "C for 90 min in the presence of sodium carbonate, potassium carbonate or potassium t-butoxide resulted in the formation of a mixture of two monobenzyl and, because of tautomerization, two dibenzyl derivatives, the relative amounts of each compound being dependent on the base used. The four alkyl derivatives could be readily separated by GC, and their retention indices are shown in Table 2. The molecular weight of each compound was confirmed by GCMS. It was not possible to determine the substitution patterns from the mass spectra. When the same series of reactions was carried out using benzene as solvent and after adding 18-crown-6 ether as catalyst, it was found that the dibenzyl derivative with the gxeater retention time was formed almost exclusively (>goo/,) when potassium t-butoxide was the base; whereas the crown ether had little effect on the @ Heyden & Son Ltd, 1979
Retention index'
2290 2500 2780 2970
Monobenzyl Monobenzyl Dibenzyl Dibenzyl
a The retention indices were measured on a 1.8 m x 4 mm i.d. glass column packed with 3% OV-l on Chromosorb W-HP (100/120 mesh). The oven temperature was 250 "C and the carrier (nitrogen) flow was 60 ml min-'.
relative amounts of products from the reactions using sodium or potassium carbonate. ' T h e reaction of tiamenidine with benzyl bromide in the presence of 18-crown-6 ether proceeds equally well in hexane and this solvent was preferred in the final method. Crown ethers have been used previously to prepare phenacyf and pentaflu~robenzyl~ esters for chromatography. Generally the esters are obtained in qualitative yields (>97%) free from by-products. The extension, reported here, of this type of reaction to the formation of N-alkyl derivatives could well have wide application.
Mass spectra The mass spectra of the dibenzyl derivatives of tiamenidine and desmethyltiamenidine formed in the presence of 18-crown-6 ether are shown in Figs. 2 and 3. As the molecular ions are small, the [M-C1]+ ions, which account for approximately 25% of the total ionization in each case, are monitored in the method.
Accuracy and precision The accuracy and precision of the method was determined by analysing plasma to which had been added tiamenidine hydrochloride over the range 0.2 ng m1-I10 ng rn1-l. Each sample was analysed six times and the results obtained are shown in Table 3. Comparison of 1001
I
m/z
Figure 2. Mass spectrum of the dibenzyl derivative of tiamenidine formed in the presence of 18-crown-6 ether.
BIOMEDICAL MASS SPECTROMETRY, VOL. 6, NO. 1. 1979 29
T. A. BRYCE AND J. L. BURROWS
91
Table3. Analysis of plasma after the addition of various amounts of tiamenidine hydrochloride 3L6
rl
* I
1 " " I " ~ ~ l
200 m /I
I,
,!, , ,
,,
,
300
,,
1 a
Tiamenidine hydrochloride added (ng rn1-l)
Tiamenidine hydrochloride found Mean SD (6 determinations) (ng m1-l)
0 0.21 0.52 1.01 2.04 5.20 10.3
0.06* 0.04 0.25* 0.03 0.55*0.04 1.11*0.19 2.02*0.17 4.98*0.31 10.8 *0.6
*
Recovery of tiamenidine hydrochloridea
1%)
90* 12 94* 7 104*17 96* 8 95+ 6 104* 6
After subtraction of blank value of 0.06 ng ml-'
Figure 3. Mass spectrum of the dibenzyl derivative of desrnethyltiamenidine formed in the presence of 18-crown-6 ether.
the mean determined tiamenidine concentrations with the actual tiamenidine levels shows that the accuracy is within *looh.The precision was found to be better than *20% for levels of 1 ng ml-' and below, and better than *lo% for levels greater than 1 ng ml-'.
Sensitivity When blank plasma samples were analysed, a small peak was always observed [Fig. 2(A)] at the retention time
of dibenzyltiamenidine. It has not been established whether this peak represents a co-extracted compound or is derived from dibenzyltiamenidine washed out from the syringe, column or interface. However, the average amount of this peak was found to be (Table 3) equivalent to 0.06 ng ml-' tiamenidine hydrochloride. Although this limits the sensitivity of the method, satisfactory values of accuracy and precision were obtained for concentrations of 0.2 ng m1-l.
REFERENCES 1. F. Kersting,Arzneim.-forsch.23,1657 (1973);E. Lindnerand J. Kaiser, Arch. Int. Pharmacodyn. Ther. 221, 305 (1974). 2. G. H. Draffan, R. A. Clare, S.Murray, G. D. Bellward, D. S. Davies and C. T. Dollery, in Advance in Mass Spectrometry in Biochemistry and Medicine, Vol. 2, ed. by A. Frigerio, p. 389, Spectrum Publications, New York (1977). 3. P.-0. Edldnd and L. K. Paalzow. Acta Pharmacol. Toxicol. 40-, 145 ( 1977). 4. H.-W. Fehlhaber, K. Metternich, D. Tripier and M. Uihlein, Biomed. Mass Spectrom, 5, 188 (1978).
30 BIOMEDICAL MASS SPECTROMETRY, VOL. 6, NO. 1, 1979
5. H. Stahle and K.-H. Pook, LiebigsAnn. Chem. 751, 159 (1971). 6. H. D. Hurst, M. Milano, E. J. Kikta Jr, S. A. Connelly and E. Grushka, Anal. Chem. 47, 1797 (1975). 7. 8. Davis, Anal. Chem. 49,832 (1977).
Received 26 July 1978 @ Heyden & Son Ltd, 1979
@ Heyden & Son Ltd, 1979
