| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
University of Helsinki, Department of Chemistry, Laboratory of Analytical Chemistry, Helsinki 00010, Finland.
2
Kimron Veterinary Institute, National Residue Control
Laboratory, PO Box 12, Beit Dagan 50250, Israel.
a Author for correspondence. Fax 972-3-9681692; e-mail ssoba_vs{at}netvision.net.il
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: The three components were isolated by preparative chromatography and their identities verified by thin-layer chromatography, HPLC, mass spectrometry, nuclear magnetic resonance spectroscopy, and melting point determination. The gentamicins were extracted from the biological matrix by use of Tris buffer and polymer phase solid-phase extraction. Derivatization was carried out in the solid-phase extraction cartridge with 1-fluoro-2,4-dinitrobenzene. The 2,4-dinitrophenyl derivatives were separated with reversed-phase HPLC and quantified by the ultraviolet absorbance at 365 nm.
Results: The detector response was linear from the limit of quantification to 50 mg/L for the individual components. The limit of quantification was 0.07 mg/L for gentamicin C1 and 0.1 mg/L for gentamicins C2 and C1a. The recovery of the gentamicin components was 72% from plasma and 98% from urine. The method was validated for human and dog plasma and urine.
Conclusions: The method was repeatable and enabled the analysis of gentamicins C1, C1a, and C2 in plasma and urine in concentrations covering the therapeutic range of the drug, thus being suitable for therapeutic drug monitoring and pharmacokinetic studies.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). The C2 component consists of two
stereoisomers (C2 and C2a).
The structural difference between the gentamicin components is
comparable to the difference between gentamicins and other
aminoglycosides such as tobramycin and netilmicin. Despite the wide
therapeutic use of gentamicin for decades and the importance of
gentamicin in clinical practice, chromatographic methods able to
determine the individual components in biological fluids are few and
are limited in performance (2). Consequently, the accuracy
of the methods generally used for measuring total gentamicin
concentration, such as immunoassays and bioassays, has not been
confirmed as a function of gentamicin component ratio.
|
Gentamicin use is associated with severe side effects, the most important of which are nephrotoxicity and ototoxicity (1). Animal data indicated that the different gentamicin components differed in their toxicities (3), but human data were insufficient to confirm this (4). Because of its toxicity, gentamicin use is subject to therapeutic drug monitoring, especially in patients with renal impairment. The research concerning disposition of the individual gentamicin components has been practically nonexistent, apparently because of the lack of suitable analytical methods and calibrators. Gentamicin C1 was reported to have different pharmacokinetics than the gentamicin complex when gentamicin C1 was given separately to patients (4). Subsequently, results from a study using an HPLC method able to separate gentamicin C1 from the two other components contradicted these results (5). Decisive determination of the pharmacokinetics of the individual gentamicin components may have important therapeutic and toxicological implications.
Currently, the set limits for the component ratio, as defined in the US Pharmacopoeia (6), are 25–50% for C1, 10–35% for C1a, and 25–55% for the sum of C2 and C2a. The British Pharmacopoeia (7) limits are 25–50% for C1, 15–40% for C1a, and 20–50% for the sum of C2 and C2a. However, the official methods for analysis of the gentamicin component ratio in drug standards have been criticized for failing to indicate the actual composition of a gentamicin complex (8). Consequently, the component ratio in commercial gentamicin preparations varies remarkably and is generally unknown (8). The importance of the variability in the component ratio and its implications to gentamicin toxicity remain to be established (9).
The chromatographic methods that separate gentamicins C1, C1a, and C2 lack sufficient sensitivity and accurate identification and/or quantification to measure therapeutically relevant concentrations of the individual components in plasma and urine (2). Therefore, fluorescence polarization immunoassays and radioimmunoassays currently are the principal methods used in therapeutic drug monitoring and pharmacokinetic studies of gentamicin, but microbiological assays have also been used (9). These methods lack the ability to identify and measure separately the three components. It also is not clear whether these methods are able to measure with equal accuracy the concentrations of the individual components, hence producing a potential bias in the total gentamicin concentration. In a recent study of a homogeneous immunoassay for gentamicin, this issue was addressed (10). Notwithstanding that the authors claimed similar response and that absolute recovery was not determined, the results indicated that differences up to 32% between the reactivities of the antibodies to the gentamicin components may exist. Remarkable intersubject variability has been reported for immunoassays in general, and the accuracy of the results has been questioned, especially at low concentrations (11). Thus, development of a chromatographic method for reliable detection, identification, and quantification of the three major gentamicin components in body fluids in the therapeutic range (0.1–20 mg/L) is essential for further research concerning gentamicin pharmacokinetics and toxicity.
The present study describes the development of a liquid chromatographic method able to determine concentrations of gentamicins C1, C1a, and C2 in plasma and urine at relevant concentrations for therapeutic drug monitoring and pharmacokinetic studies.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
instrumentation
A Hewlett-Packard H-P 1100 low-pressure mixing gradient HPLC
equipped with a diode array ultraviolet-visible detector, autosampler,
and column oven was used. An Esquire (Brucker and Hewlett-Packard) ion
trap mass spectrometer with an atmospheric pressure chemical ionization
(APCI) interface was utilized for mass spectrometric determinations. A
Varian VXR-300S nuclear magnetic resonance (NMR) spectrometer was used
for measuring the proton NMR.
hplc procedure
Gentamicins C1, C1a,
and C2 were determined as their 2,4-dinitrophenyl
derivatives using a mobile phase of 680 mL/L acetonitrile-320
mL/L Tris buffer (8.3 mmol/L, titrated to pH 7.0 with HCl) at a
flow rate 1.2 mL/min. A SymmetryTM
C18 reversed-phase column (100 x 4.6 mm;
3.5 µm bead size; Waters) connected to a C18
precolumn was used for separation. The injection volume was 20 µL,
the column temperature was 25 °C, and the chromatographic eluent was
monitored at 365 nm.
preparative separation of the gentamicin components
The gentamicin C1,
C1a, and C2 components were
separated from gentamicin sulfate (lot no. 42H0610; Sigma) in a silica
column according to the method described by Claes et al (8).
The column was prepared from a silica (100 g) slurry in
isopropanol-chloroform-170 mL/L ammonium hydroxide (1:2:1, by
volume). The column was eluted with isopropanol-chloroform-250 mL/L
ammonium hydroxide (1:2:1, by volume), and 10-mL fractions were
collected. The elution was monitored using the thin-layer
chromatographic method described by Wilson et al.
(12). The solvent system for the thin-layer chromatography
consisted of a lower phase of methanol-chloroform-250 mL/L ammonium
hydroxide (1:1:1, by volume), and the gentamicins were visualized with
ninhydrin spray (0.1 g ninhydrin in 100 mL of 950 mL/L
n-butanol-50 mL/L glacial acetic acid). The appropriate
fractions were combined and evaporated to dryness. The residue was
redissolved in methanol and evaporated under reduced pressure
to constant weight. The purities of the component calibrators were
assessed by HPLC, thin-layer chromatography, NMR, and melting point.
The proton NMR was measured, and the chemical shifts are all reported
relative to tetramethylsilane. The antimicrobial potency of the
calibrators was evaluated using Bacillus subtilis as test
organism in an agar diffusion disk assay.
preparation of calibrators and calibration curves
We prepared the calibrators by diluting each of the gentamicin
components in deionized water. The stock solutions were added
to plasma and urine to provide concentrations of 0, 0.1, 0.2,
0.4, 1, 5, 10, 15, 20, and 50 mg/L of the individual gentamicin
components. The calibrators in water, plasma, and urine were
derivatized in the SPE cartridge as described below. The calibration
curves were obtained by plotting the peak areas as a function of the
respective concentrations for each analyte and calculating the linear
regression. Recoveries from plasma and urine were determined by
comparison of samples to which the components had been added to aqueous
calibrators, all prepared using SPE. The relative recovery of the SPE
and derivatization was determined by comparing responses of aqueous
calibrators prepared by SPE to aqueous calibrators evaporated to
dryness and derivatized in a vial with 0.3 mL of the derivatization
solution (100 °C for 1 h).
derivatization reagent
The derivatization reagent consisted of 0.5 mL of 0.17 mol/L Tris
buffer (pH 12.0), 0.5 mL of water, 2 mL of acetonitrile, and 50 mg of
FDNB in 0.2 mL of acetonitrile. The reagent, except for the FDNB
component, was stored in 3°C. The FDNB fraction was added to the
reagent before use.
preparation of plasma and urine samples
The SPE cartridge was conditioned with 1 mL of methanol followed
by 1 mL of 0.17 mol/L Tris buffer at pH 10.0. Five milliliters of Tris
buffer (0.17 mol/L Tris, titrated to pH 12.0 with NaOH) was added to
1.0 mL of plasma or urine, vortex-mixed, and charged on the cartridge.
The flow through the cartridge was kept at <0.3 mL/min. The cartridge
was washed with 2 mL of 0.17 mol/L Tris buffer (pH 10) and dried; 300
µL of the derivatization mixture was then applied on the SPE
cartridge. The top of the cartridge was closed with a conical
polypropylene pipette tip (5–200 µL) to reduce evaporation. The
cartridge was then incubated for 1 h in a laboratory oven at
100 °C. The derivatized gentamicin was eluted with 5 mL of
acetonitrile by gravity into the tube holding the cartridge. The eluate
was evaporated to dryness, redissolved in 300 µL of acetonitrile, and
transferred to an autosampler vial. The blood samples from
gentamicin-treated dogs or humans were collected in
heparin-containing test tubes, and plasma was separated with
centrifugation. The plasma samples were then stored at -30 °C until
analysis as described above.
mass spectrometry
The preparatively separated gentamicin C1,
C1a, and C2 components were
identified using APCI and ion trap mass spectrometry. The
analytes were dissolved in a solution of methanol-0.05 mol/L ammonium
acetate (1:1, by volume) and injected directly (by syringe pump) into
the liquid chromatography-mass spectrometry interphase. Positive ions
were detected. The APCI heater temperature was set at 400 °C. The
capillary exit was maintained at 100.0 V, skimmer 1 maintained at 30 V,
and the cutoff was at 50 m/z. Nitrogen was used as a sheath
gas, and helium was the collision gas.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. The mass spectrum of gentamicin C1 is
presented in Fig. 2
.
|
|
Proton NMR gave the following results:
For C1a (D2O,
): 1.06 (s, 3H), 1.15–1.4 (m, 2H), 1.5–1.72 (m, 2H),
1.78–1.9 (m, 2H), 2.49 (s, 3H), 2.6–2.9 (m, 6H), 3.05–3.3 (m, 2H),
3.35–3.45 (m, 2H), 3.7–3.95 (m, 3H), 4.85–4.95 (m, 2H)
For C2 (D2O, ):
0.86–0.91 (d, 3H), 1.04 (s, 3H), 1.2–1.5 (m, 2H), 1.55–1.7 (m, 2H),
1.75–1.85 (m, 2H), 2.44 (s, 3H), 2.48–2.52 (d, 1H), 2.7–2.9 (m, 4H),
3.15–3.3 (m, 2H), 3.38–3.45 (m, 1H), 3.45–3.58 (m, 1H), 3.7–3.8 (m,
1H), 3.97 (s, 2H), 4.90–5.0 (m, 2H)
For C1 (D2O, ):
0.89–0.92 (d, 3H), 1.07 (s, 3H), 1.2–1.55 (m, 2H), 1.56–1.7 (m, 2H),
1.75–1.9 (m, 2H), 2.27 (s, 3H), 2.47 (s, 3H), 2.48–2.58 (d, 1H),
2.55–2.65 (m, 1H), 2.7–2.92 (m, 3H), 3.15–3.3 (m, 2H), 3.25–3.36
(m, 1H), 3.7–3.78 (m, 1H), 3.78–3.85 (m, 1H) 3.95 (s, 2H), 4.95–5.1
(m, 2H)
where is the chemical shift in parts per million downfield
from the tetramethylsilane standard, s is a singlet, d is a doublet,
and m is a multiplet.
The melting point of C1 was 92–97 °C, and that of C2 was 105–112 °C. The melting point of gentamicin C1a was not determined because there was no reference value.
The purified components did not differ in their antimicrobial activity, as reported previously (13). In addition, the antimicrobial activity of the individual components was equal to that of the reference gentamicin mixture.
chromatographic separation
Chromatograms of the separation of the mixture of gentamicins in
dog plasma and a blank chromatogram are presented in Fig. 3
. The chromatographic peaks were identified using single
component calibrators obtained as described above. The elution order of
the gentamicins was C1a, C2
isomers, and C1. Only one peak was obtained after
derivatization of each gentamicin component calibrator, indicating that
derivatization was complete. Separation was obtained only within a
narrow range of mobile phase constituent concentrations. When the
acetonitrile concentration was 50 mL/L higher, the separation was lost,
whereas when it was 50 mL/L lower, the retention times increased and
the peaks became unacceptably wide.
|
method validation
The method was validated for human and dog plasma and urine. The
precision, linearity, and recovery of the method were evaluated. For
the interassay CVs, 20 repetitive analyses were performed within 2
months on three different concentrations (0.5, 5, and 50 mg/L). The
calculated CVs and recoveries for each analyte are presented in Table 2
. The limit of quantification of the method, defined as
nine times the background noise, was 0.1 mg/L for gentamicins
C2 and C1a, and 0.07 mg/L
for gentamicin C1 from plasma and urine samples.
The plasma and urine calibration curves were linear with the following
equations: gentamicin C1, y =
340x (r2 = 0.999);
gentamicin C2, y =
169x (r2 = 0.998); and
gentamicin C1a, y =
183x (r2 = 0.996) for
plasma; and gentamicin C1, y =
470x (r2 = 0.999);
gentamicin C2, y =
235x (r2 = 0.998); and
gentamicin C1a, y =
245x (r2 = 0.998) for
urine. Amoxicillin, ampicillin, neomycin, tobramycin, streptomycin,
sulfadiazene, sulfamethoxazole, oxytetracycline, and doxycycline were
added to plasma samples for evaluation of possible interference. No
interference from these drugs or common plasma and urine components was
detected in >200 samples analyzed.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Gentamicin does not possess a chromophore in its structure and must be derivatized to enable ultraviolet detection. After preliminary trials with various reagents, we chose FDNB derivatization for the gentamicin derivatization. The 2,4-dinitrophenyl derivatives are very soluble in organic solvents and thus easy to elute from the SPE cartridge. The specific stability of the Oasis cartridges over a wide pH range enabled the use of basic buffers, which was essential for gentamicin retention in the cartridge and for on-column derivatization. The SPE material was stable in the temperatures used and allowed optimization of the derivatization yields. The SPE and on-column derivatization allowed the use of much higher plasma volumes than in earlier studies (generally <500 µL) (2). Consequently, assay sensitivity could be enhanced. The present sample preparation and derivatization procedure is also fast and simple to perform and allows throughput of hundreds of samples as required in pharmacokinetic studies. The high stability of the dinitrophenyl derivatives, superior to most other derivatization reagents, allows preparation of large numbers of samples simultaneously. A single analysis could be performed in less than 3 h.
Our results show that FDNB-derivatized gentamicin components produce different detector responses. o-Phthaldialdehyde (OPA) with mercaptan has been widely used for the derivatization of gentamicin (2). In preliminary studies, we found that OPA postcolumn derivatization with mercaptoethanol did not allow separation of the three components and that precolumn OPA derivatization did not produce the necessary assay sensitivity, probably because of the instability of the derivatives. The British Pharmacopoeia and US Pharmacopoeia claim that when precolumn OPA derivatization is used, the individual peak areas, as percentages of the total peak area, can be used to quantify the gentamicin component ratio. Both the British and the US Pharmacopoeia methods for determining the ratio of gentamicin components have been shown to be erroneous (8). It seems likely that the inaccuracy results from variations in detector response generally believed to correlate with the number of derivatized groups. Consequently, gentamicin C1 should have a smaller response than the two other components because it has one less free amine and OPA reacts only with primary amines. Therefore, the differences in detector responses obtained here for the three gentamicin components substantiate the prerequisite that the components be analyzed independently and that the use of individual component calibrators is necessary for adequate quantification of total gentamicin.
The chemical differences between the gentamicin components are very small, and sufficient separation of the components is difficult to achieve. Barends et al. (14) reported derivatization of gentamicins with FDNB, but only C1a was separated from the two other components. In the present study, the dinitrophenyl derivatives of gentamicins C1a, C2, and C1 were sufficiently separated to allow quantification of the individual compounds. This was apparently attributable to differences in the column specifications because we could also achieve the same separation with ammonium acetate buffer (1.0 mmol/L at pH 4.2). The gentamicin C2a enantiomer eluted together with gentamicin C2 on a standard C18 column.
It must be emphasized that the gentamicin content in commercial calibration materials is given as the antimicrobial potency of an unknown mixture of at least three components. This severely invalidates accurate quantification based on use of these calibrators. Method development also becomes questionable because calibration curves cannot be prepared without the use of calibrators for the individual component.
applicability of the method
The appropriateness of the present method was assessed in a
study performed to determine the pharmacokinetics of the gentamicin
components in dogs. To our knowledge, this is the only study describing
the individual pharmacokinetics of the three gentamicin components. The
results indicated that the pharmacokinetic characteristics of the
components were significantly different. The results of the study will
be reported elsewhere (15). Currently, the method is
used in a pharmacokinetic study of gentamicin components in human
patients. Representative plasma concentration-time curves of the three
gentamicin components from a patient are presented in Fig. 4
.
|
The present method can be used to evaluate the accuracy of commonly used nonchromatographic gentamicin assays and to determine the component ratio in unknown samples. In light of the results of this study, it appears increasingly important to establish the effect of variations in the ratios of the components on the accuracy of methods unable to separate the components.
To our knowledge, this is the first time that FDNB was used for derivatization of amines in a SPE cartridge. The present extraction and derivatization technique may facilitate sample preparation of a wider class of hydrophilic amines that are difficult or impossible to extract or concentrate from biological matrixes as the parent compounds.
In conclusion, the present method allows fast analysis of gentamicins C1a, C2, and C1 in plasma or urine. It is the first method able to separate the three major gentamicin compounds that has been applied to pharmacokinetic studies. Sample preparation using SPE and on-column derivatization with FDNB allowed use of large sample volumes and produced stable, concentrated gentamicin dinitrophenyl derivatives. The need for quantification of the individual components to determine the total gentamicin concentration was emphasized. The method can be used as a reference method for performance characterization of nonchromatographic methods.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
), C2
(
), and C1 (
), respectively, demonstrating the
differences in the pharmacokinetics of the three components.