| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
a Author for correspondence. Fax 49-6131-176619; e-mail lichtwal{at}mail.uni-mainz.de.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We here present a method that permits injection of an unprocessed urine sample and the simultaneous measurement of E, NE, dopamine (DA), metanephrine (M), normetanephrine (NM), 3,4-dihydroxyphenylacetic acid (DOPAC), and homovanillic acid (HVA), as well as serotonin (5HT) and 5-hydroxyindole-3-acetic acid (5HIAA). System-integrated sample processing was achieved by the use of a restricted access silica precolumn device. The precolumn is coupled by an electrically driven valve to an analytical column on which the analytes were chromatographed. Ion-pair reversed-phase chromatography was used as separation mode, and was followed by postcolumn derivatization and fluorometric detection.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. We used the model 2700 Solvent Delivery System
(Bio-Rad), which consisted of two dual piston HPLC pumps (pumps
1 and 2 in Fig. 1
). The Hitachi reaction pump (model 655A-13; Merck)
has three parallel pump heads for feeding three reagents, two were used
as pumps 3 and 4. Injections were made with a Rheodyne valve (model
7125) equipped with a 100-µL sample loop. The LiChrospher RP-18 ADS
precolumn (particle size, 25 µm; column size, 25 x 4 mm i.d.;
Merck) was connected via an electrically driven six-port switching
valve (WE C6WK; Valco Europe, distributed by Merck) to the analytical
reversed-phase column (LiChrospher 100 RP-18; particle size, 5 µm;
column size, 125 x 4 mm i.d.; Merck). The coils in the
derivatization unit were made of Teflon (0.33 mm i.d.). Coils a and b
were 2 and 7 m long and were heated in a column oven (LaChrom
L7350; Merck) at 95 °C. Coil c was 1.2 m long and was placed
together with the analytical column in a thermostated water bath (type
D1+G; temperature range, -10 to 100 °C; Haake) at 10 °C. The
fluorescence emissions of the analytes were measured at 480 nm
emission, with excitation at 350 nm, by a Hitachi fluorescence
spectrophotometer (model F-1050; Merck) equipped with a 12-µL flow
cell (cat no. 050-1322). The High Resolution Liquid
Chromatographic System interface (model 822; Bio-Rad) controlled
the switching valve and pumps 1 and 2. In addition, it transferred the
data from the detector to the Bio-Rad software (Series 800 High
Resolution Liquid Chromatographic System, Ver. 2.30.1a), which
monitored the chromatograms and determined the areas under the detected
peaks.
|
chemicals
DA was obtained from Serva; DOPAC and HVA were purchased from
ICN; and 5HIAA, M, NE, NM, and 5HT were from Sigma-Aldrich. E and the
other chemicals were at least of analytical reagent grade and were
supplied by Merck. meso-1,2-diphenylethylenediamine (DPE)
was synthesized by the procedure of Irving and Parkins (11).
solutions
The mobile phase for both columns was a buffer solution
containing 0.1 mol/L sodium dihydrogen phosphate monohydrate, 5 mmol/L
sodium octyl sulfate, and 0.1 mmol/L sodium azide. The pH was adjusted
to 2.5 with orthophosphoric acid (200 mL/L). The oxidizing
reagent was an aqueous solution composed of 20 mmol/L sodium periodate
and 6 mmol/L potassium hexacyanoferrate (III). The fluorescence reagent
was 700 mL/L ethanol solution containing 60 mmol/L DPE and 0.3 mol/L
sodium hydroxide. The deionized water used for all methods was purified
by the Millipore reagent grade water system (MilliQ ZFMQ 23004).
All solutions were degassed for 5 min in an ultrasonic bath before use.
The stock mixture used for calibration contained
10-4 g/L DA, E, and NE, and
10-3 g/L M, NM, DOPAC, HVA, 5HT, and 5HIAA. All
calibrators were constituted from 1 g/L stock solutions in mobile phase
buffer.
urine specimens
Urine specimens (24 h) were collected and acidified with 2.5
mol/L sulfuric acid containing 100 g/L glycine (10 mL/L urine).
Aliquots (10 mL) were stored at -20 °C. Before injection, the urine
was centrifuged at 1200g for 10 min at room temperature.
To assess imprecision specimens from 20 normotensive and hypertensive adult individuals were measured. The samples used had previously shown no pathological catecholamine concentrations according to the method of Kringe et al. (12).
Aliquots (10 mL) from 113 individual nonpathological 24-h urine samples were pooled for use as the in-house control.
Three single specimens from patients with pheochromocytoma confirmed by surgery were measured.
For method validation, two control urines were obtained from Chromsystems, one with physiological (control urine I) and the other with pathological concentrations (control urine II) of the relevant analytes and contained the following concentrations: control urine I, 66 µg/L NE, 13 µg/L E, 265 µg/L NM, 129 µg/L M, 180 µg/L DA, 3.4 mg/L DOPAC, 4.2 mg/L HVA, 141 µg/L 5HT, and 5.5 mg/L 5HIAA; control urine II, 213 µg/L NE, 50 µg/L E, 1.1 mg/L NM, 301 µg/L M, 525 µg/L DA, 11 mg/L DOPAC, 15 mg/L HVA, 0.9 mg/L 5HT, and 31 mg/L 5HIAA.
procedure
Sample was applied when the valve was in position A (Fig. 1
).
The mobile phase, delivered by pump 1 at a flow rate of 0.3 mL/min,
eluted mainly the high-molecular weight and hydrophobic components of
the sample from the ADS precolumn to waste. Simultaneously, the
analytes were retained on the hydrophobic bonded phase of the sorbent.
Five minutes after sample injection, the six-port valve rotated 60°
to position B, coupling the precolumn to the analytical column, and the
mobile phase, delivered by pump 2 at a flow rate of 0.8 mL/min, eluted
the analytes from the precolumn. The precolumn and the analytical
column were coupled for 11 min to allow transfer of the analyte
fraction onto the analytical column, after which the valve switched
back to position A and separation on the analytical column continued.
Methanol (200 mL/L) was added to the mobile phase 110 min after
injection to accelerate the elution of the last analyte (5HT). During
the whole procedure, pumps 3 and 4 of the derivatization unit
introduced the oxidizing and fluorescence reagents to the column eluate
at a flow rate of 0.15 mL/min. The operating sequence is listed in
Table 1
.
|
calculations
To calculate analyte concentrations, the areas under the peaks of
unknown samples were related to the peaks of the stock mixture used as
calibrator. The linear regression data were processed by the program
Microcal Origin, Ver. 4.10 (Additive).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. Nine compounds could be separated within 150 min, including
sample pretreatment. The shift of the baseline at 110 min is the result
of the isocratic introduction of the mobile phase containing 200 mL/L
methanol.
|
sample pretreatment
The ADS column, which permit a system-integrated sample
processing, was developed by Boos et al. (13); to our
knowledge, the present study was the first time it was used to separate
catecholamines and their metabolites within a single analysis. This
porous alkyl-diol silica (LiChrospher RP-18 ADS) consists of a
hydrophilic and electroneutral external particle surface and a
hydrophobic reversed-phase internal surface. These bimodal
chromatographic properties allow retention of hydrophobic low-molecular
weight analytes by classical reversed-phase chromatography exclusively
at the hydrophobic pore surface. Macromolecular constituents of the
sample matrix are size-excluded by 6 nm pores and eluted into the
waste.
Online analysis was performed by coupling of the ADS-precolumn and the analytical column via an electrically driven six-port valve. The switching of the valve to position B ended sample pretreatment on the precolumn and coupled it with the analytical column. The switching point was programmed at 5 min after injection. At this time point, most of the sample matrix, monitored by a detector (wavelength 280 nm) in a previous test analysis, had eluted from the ADS; however, NE, the first analyte to elute from the analytical column, was still retained. The change in flow direction (back flush) after the valve was switched achieved a transfer of the concentrated analytes as a single band to the analytical column. This was monitored by an in-line detector (wavelength, 280 nm) between the precolumn and the analytical column. Eleven minutes after the valve switched to position B, it switched back to position A.
separation of the analytes
The catecholamine-related compounds contain amino, carboxyl, and
alcohol groups in their molecular structure. For the simultaneous
separation of such analytes, ion-pair reversed-phase HPLC is the method
of choice. Separation was achieved with the mobile phase described with
5 mmol/L sodium octyl sulfate added as the ion-pair reagent.
Concentrations of sodium octyl sulfate as high as 5 mmol/L were
essential for maintaining constant retention times, especially for
late-eluting analytes.
In addition to the LiChrospher 100 RP-18 column, we tested three other C18 reversed-phase columns: one column packed with Nucleosil 100–5 C18 (200 x 4 mm; Machery & Nagel), one column "for catecholamine analysis", but not otherwise specified (100 x 4 mm; Chromsystems), and one column packed with SilicaROD RP-18 (100 x 4 mm; Merck, Darmstadt, Germany). The last column showed a pressure at the pumps that was reduced to 27%, but the LiChrospher column gave the best resolution.
A constant column temperature is important for the analysis. A linear relationship between the retention times of the catecholamine-related compounds and the column temperature was found. At low temperatures, resolution of the peaks was higher, retention times were increased, and peak broadening was reduced. The decision was made to use the analytical column at 10 °C.
postcolumn derivatization
Postcolumn derivatization was based on the method by Jeon et al.
(14). In this reaction, the catecholamine-related compounds
are first oxidized with periodate and potassium hexacyanoferrate to the
corresponding o-quinones. Those activated molecules form
with the introduced DPE via azomethines to 2-phenylbenzoxazole
derivatives, which show fluorescence at 480 nm after excitation
at 350 nm (15). We found that 5HT and 5HIAA also show
fluorescence under these conditions, although they could not be
oxidized to o-quinones, which are necessary intermediate
compounds for the formation of 2-phenylbenzoxazole derivatives. Their
signals were reduced to 13% compared with that of E on a molar basis.
Without the oxidizing reagent and DPE, none of the analytes, in
physiological concentrations, fluoresce at the 480 nm with the chosen
detection gain.
For the derivatization unit, we used different materials than did Jeon et al. (14). We tested the length of the reaction coils and found that lengths of 2 m for the oxidation and 7 m for the fluorescence reaction offered the best compromise between peak height and peak width. The coils were heated in a column oven at 95 °C. The apparent temperature of the eluate stream was 48 °C. E, with the lowest concentration of the nine analytes in urine under physiological conditions, produced maximum peaks at this temperature. A DPE concentration of 0.06 mol/L and an apparent pH 6.0 in the fluorescence reaction coil, which is achieved with 0.3 mol/L sodium hydroxide in the fluorescence reagent, were the most favorable for the condensation and cyclization reactions. Before reaching the detector, the eluate passed through the 1.2-m coil in a cooling bath (10 °C) and was cooled to 19 °C.
validation
Identification.
We identified the peaks in urine by monitoring
the coincidence of the retention times and the increases in peak
heights when stock solutions of the calibrators were cochromatographed
with the urine samples. The analytes showed no peaks when DPE and the
oxidizing reagents were omitted from the solutions for the
derivatization.
Linearity and sensitivity.
Injections of stock solutions in
the concentration range 1 x 10-6. to 1
x 10-2 g/L indicated that the monitored signals were
linear for each analyte. The linear regression data are
presented in Table 2
. The slopes of the regression lines are the measures for the
analytical sensitivity.
|
Precision.
Intra- and interassay imprecision as determined
from analysis of the stock mixture and control urines I and II, is
summarized as CVs in Table 3
. The concentrations of the analytes in each of the control
urines are listed in Materials and Methods.
|
Recovery and agreement with expected values for control
urines.
The analytical recoveries are shown in Table 4
. For determination of the matrix-independent recovery, the
stock mixture was chromatographed 10 times with ADS sample pretreatment
and 10 times without ADS. The values without pretreatment were set as
100% and the values with pretreatment were related to them. Therefore,
the matrix-independent data describe the influence of sample
pretreatment on the recovery for the stock mixture. For the
matrix-dependent recovery, known amounts of stock solutions at three
different concentrations were added (each three times) to control urine
I and analyzed. The concentrations of analytes in the stock solutions
and urine were also measured separately in single analyses. The
results of the cochromatographed samples were related to the sum of the
single samples. Therefore, the matrix-dependent data describe the
influence of the sample matrix on recovery. In addition, we compared
the values of control urine I specified by the manufacturer with the
concentrations determined (n = 20). The deviations from the
expected values are also summarized in Table 4
.
|
Comparison of methods.
The measured values for E and NE in 27
urine samples from different individuals were compared with those
determined with the method of Kringe et al. (12). Their
HPLC procedure included extraction by aluminum oxide, reversed-phase
separation, and fluorescence detection of the trihydroxyindole
derivatives, and determined only E and NE. The correlation of E and NE
measured with both methods is 97.5% and 96.6%, respectively (Fig. 3
).
|
Detection limits.
The detection limits for the analytes at a
signal-to-noise ratio of three were based on the linear dilution of
their stock solutions (Table 4
).
Values in subjects with and without pheochromocytoma.
Each
analyte was determined in single 24-h urine samples collected from 20
normotensive and hypertensive adult individuals (Table 5
). In addition, we measured a pooled urine that contained
aliquots of 113 24-h urines. Previously, the E and NE in each of the 20
and the 113 samples were measured by the method of Kringe et al.
(12) and showed E and NE values that were not increased.
With the new method, the pooled urine was assayed 15 times. The means
of the obtained concentrations were related to the average volume of
the 113 urines. Those results are near the mean values for the 20
individuals and are shown in Table 5
contrasted to the results from 3
patients with confirmed pheochromocytoma. Patient 2 is one case of a
group whose in 24-h urines showed increased E but NE within the
reference interval. Patient 3 is a single case with additional
high 5HT values.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Simple system-integrated sample processing was achieved with a
restricted access precolumn (ADS). When column-switching techniques
were used, the analytes were eluted from the ADS precolumn to the
analytical column without dilution and external transfer. This is
documented in the nearly quantitative matrix-independent recovery
(Table 4
). The size exclusion of the precolumn eliminated many of the
compounds that interfere with the separation or contaminate the
analytical column. This effect in combination with the postcolumn
derivatization leads to a selective analysis.
The compounds were separated by means of ion-pair reversed-phase chromatography. The conditions for the separation were similar to those mentioned in the literature (9)(16)(17). A C18 column was used as stationary phase, and a buffer at pH 2.5 with the ion-pair reagent sodium octyl sulfate and an organic modifier was used as the mobile phase. Contrary to the specifications of several investigators (10)(14)(16)(17), we could not achieve constant retention times with concentrations of the ion-pair reagent lower than 5 mmol/L.
The postcolumn derivatization was based on the method of Jeon et al. (14). However, we used reaction coils with smaller internal diameters, which maintained the high pressure during derivatization and prevented the generation of bubbles, especially in the heated areas. In addition, the use of coils with smaller internal diameters reduced peak broadening. Our oxidizing and fluorescent reagents were, in contrast to those of Jeon et al., twice as concentrated but were introduced at flow rates 50% lower than the flow rates recorded in that study. This caused less dilution of the separated analytes in the coils. Before the eluate reached the fluorescence detector, its temperature was reduced by a cooling coil in a thermostated bath. Jeon et al. used only an air-cooled coil. Effective cooling is important for holding the flow cell near room temperature. Furthermore, a high solvent temperature causes a decline in fluorescence intensity.
The DPE-dependent fluorescence of 5HT and its metabolite 5HIAA was
observed at 480 nm. Although they do not undergo conversion to
benzoxazole derivatives, they are suitable informative analytes in the
diagnostic field of disturbed neurotransmitter secretion (Table 5
).
More peaks are separated in the chromatograms of urine than the defined analytes. These unidentified compounds, which show DPE-dependent fluorescence, have yet to be described.
The detection limits of the method are sufficient for urinary
measurements. To date, the low concentrations of these analytes seen in
plasma samples cannot not be detected by this method. The values
for urine specimens agree with reference intervals determined with
other HPLC methods (1)(2)(18). The
values for the pooled urine sample from 113 selected normotensive and
hypertensive adults were within the ranges for the 20 subjects without
pheochromocytoma (Table 5
).
|
Acknowledgments |
|---|
|
Footnotes |
|---|
1 Nonstandard abbreviations: E, epinephrine; NE, norepinephrine; DA, dopamine; M, metanephrine; NM, normetanephrine; DOPAC, 3,4-dihydroxyphenylacetic acid; HVA, homovanillic acid; 5HT, serotonin; 5HIAA, 5-hydroxyindole-3-acetic acid; and DPE, meso-1,2-diphenylethylenediamine. 
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  R. L. Taylor and R. J. Singh Validation of Liquid Chromatography-Tandem Mass Spectrometry Method for Analysis of Urinary Conjugated Metanephrine and Normetanephrine for Screening of Pheochromocytoma Clin. Chem., March 1, 2002; 48(3): 533 - 539. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. P. Kema, W. G. Meijer, G. Meiborg, B. Ooms, P. H.B. Willemse, and E. G.E. de Vries Profiling of Tryptophan-related Plasma Indoles in Patients with Carcinoid Tumors by Automated, On-Line, Solid-Phase Extraction and HPLC with Fluorescence Detection Clin. Chem., October 1, 2001; 47(10): 1811 - 1820. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and NE (
) determined for each sample are shown.
Dotted lines represent the maximum correlation;
solid lines represent the actual regression. The
regression equation for E is: y =
0.93x + 1.0 (r = 0.975); and for NE
is: y = 0.88x + 35.3
(r = 0.966).