| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
Department of Pharmacology, University of Bergen, Armauer Hansens Hus, 5021 Bergen, Norway.
a Author for correspondence. Fax 47-55-974605; e-mail steinar.hustad{at}farm.uib.no
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: Trichloroacetic acid-treated plasma was subjected to solid-phase extraction on reversed-phase columns. The analytes were separated by micellar electrokinetic capillary chromatography in uncoated fused- silica capillaries filled with borate buffer containing 50 mmol/L sodium dodecyl sulfate, methanol, and N-methylformamide. Native fluorescence was monitored at 530 nm, using an argon laser operating at 488 nm as excitation source.
Results: The assay was linear over a concentration range of two orders of magnitude, and the limit of detection was far below physiological concentrations for all vitamers. The within-day and between-day coefficients of variation were 4–9% and 6–12%, respectively. The reference values (median, 5–95 percentiles) obtained by analyzing plasma from 63 healthy subjects were 8.6 nmol/L (2.7–42.5 nmol/L) for riboflavin, 7.0 nmol/L (3.5–13.3 nmol/L) for FMN, and 57.9 nmol/L (44.5–78.1 nmol/L) for FAD.
Conclusions: Capillary electrophoresis with laser-induced fluorescence detection allows determination of all riboflavin vitamers far below physiological concentrations. The method may become a useful tool for the assessment of riboflavin status in humans.© 1999 American Association for Clinical Chemistry
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Riboflavin deficiency has been claimed to be quite common, and high prevalences have been reported, particularly in developing countries(3)(4)(5). The riboflavin status in humans has been assessed by clinical signs, by determination of the urinary excretion of the vitamin (6), and by measurement of the activity ratio of glutathione reductase in erythrocytes (7)(8).
For the determination of vitamin concentrations, several techniques have been developed, including microbiological (9), fluorometric (9), and liquid chromatographic methods(10). Most HPLC methods have been designed for the detection of high concentrations of riboflavin in food, pharmaceutical preparations, and urine (10). Only a few have been used to measure riboflavin in whole blood (11)(12)(13)(14), serum(11)(14)(15), or plasma(16)(17)(18).
During the last decade, capillary electrophoresis (CE) methods have been developed for several biomedical applications (19). Compared with HPLC, CE generally has the advantage of small sample requirements, short separation times, and high resolution(20). The consumption of organic solvents and other chemicals is usually lower than in HPLC.
We describe here a sensitive and robust CE method for the quantification of low physiological concentrations of riboflavin, FMN, and FAD in human plasma. The method is based on micellar electrokinetic capillary chromatography combined with laser-induced fluorescence (LIF) detection.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
instrumentation
Solid-phase extraction was performed on a programmable Gilson
ASPEC sample processor (Gilson Medical Electronics) equipped with a
Rheodyne Model 7010 injector valve and a 3000-µL sample loop.
CE was performed on a Beckman P/ACE System 2210 coupled to a P/ACE LIF Detector (Beckman Instruments). The LIF detector was connected to a 20 mW 488 nm argon laser from Uniphase. The fluorescence was monitored using a 530 nm DF30 band-pass filter from Omega Optical. Beckman System Gold software (Ver. 8.10) was used for system control and data collection and processing.
capillary electrophoresis
Uncoated fused-silica capillaries (Composite Metal Services) with
external and internal diameters of 375 µm and 75 µm, respectively,
were used. The total capillary length was 67 cm, and the effective
separation length was 60 cm. The separation buffer consisted of 100
mmol/L sodium borate, pH 7.9, containing 50 mmol/L sodium dodecyl
sulfate, 100 mL/L methanol, and 20 mL/L
N-methylformamide. All solutions were passed through
0.2 µm membrane filters (Schleicher & Schuell) before use.
Capillaries were preconditioned before each electrophoretic run by
rinsing in the high-pressure mode (137 kPa) with 0.1 mol/L NaOH for
60 s and the separation buffer for 120 s. Samples were
injected for 7 s in the low-pressure mode (3.5 kPa), corresponding
to an injection volume of ~25 nL. This was followed by a second
injection of separation buffer (for 5 s), to prevent loss of
sample material. Electrophoresis was performed at 24 kV (358 V/cm) in
the normal polarity mode (positive potential at capillary inlet). The
capillary temperature was set at 29 °C. At the end of each
electrophoretic run, the capillary was rinsed (137 kPa) with separation
buffer for 60 s.
sample collection and storage
EDTA plasma was obtained by collecting blood into Vacutainer Tubes
(Becton Dickinson), giving a final EDTA concentration of 4 mmol/L. The
samples were immediately placed on ice and centrifuged
(2000g for 10 min) within 60 min. The plasma was then
processed further or stored at -80 °C until use. Sample handling
was carried out under dim light to avoid photodegradation of the
analytes.
sample processing
Plasma samples (500 µL) were mixed with 1500 µL of ice-cold
100 g/L TCA containing 15 nmol/L isoriboflavin, which served as an
internal standard. Precipitated protein was removed by centrifugation,
and 1500 µL of the supernatant was neutralized by the addition of 480
µL of 2 mol/L K2HPO4.
Solid-phase extraction using Bond Elut C-18 columns (1 mL/50 mg;
Varian) was then carried out automatically on a Gilson ASPEC sample
processor. The columns were preconditioned with 500 µL of methanol
followed by 1000 µL of 25 mmol/L sodium phosphate buffer, pH 7.0. The
columns were then loaded with 1900 µL of neutralized TCA-treated
plasma and washed with the same phosphate buffer as used for column
conditioning. The analytes were eluted with 200 µL of 600 mL/L
methanol in phosphate buffer, and the columns were then loaded with 220
µL of phosphate buffer to reduce the methanol concentration in the
pooled eluate, which was used for CE analysis. Between steps, the
columns were dried by air pressurization.
calculation of plasma concentrations
Pooled EDTA plasma containing 7.2 nmol/L riboflavin, 6.1 nmol/L
FMN, and 56.6 nmol/L FAD was supplemented with 20 nmol/L
riboflavin, 20 nmol/L FMN, and 50 nmol/L FAD and used for calibration.
The calibrator plasma was processed in the same way as the samples and
was run approximately once for every five samples. Calculation of the
vitamer concentrations was based on peak area. For each of the
vitamers, the sample concentration was calculated using the formula:
 |
where CS is the vitamer concentration in the sample, CC is the vitamer concentration in the calibrator plasma, AS is the peak area of the vitamer in the sample, AC is the peak area of the vitamer in the calibrator, AIC is the peak area of the internal standard in the calibrator, and AIS is the peak area of the internal standard in the sample.
In addition, we collected plasma in which we obtained the flavin concentration by repeated analysis against the calibrator. This plasma was stored at -80 °C and used for quality control.
stability of the internal standard and the riboflavin vitamers
Because the internal standard was added to the TCA used for plasma
protein precipitation, we investigated its stability in 100 g/L TCA. In
addition, the stability of this compound was tested at all stages of
sample preparation. The stability of riboflavin, FMN, and FAD was
investigated in EDTA whole blood, hemolyzed blood, and EDTA plasma and
at all stages of sample preparation.
In all stability experiments, samples were placed in the dark at 23 °C, and polypropylene tubes were used for sample processing and storage. To avoid interference from the interconversion of the vitamers, each of the analytes was added separately to aliquots from one donor. Isoriboflavin served as the reference calibrator for riboflavin, FMN, and FAD, and riboflavin was used as the reference calibrator for isoriboflavin. The reference calibrators were added before the elution step to correct for variations in solid-phase extraction and sample injection. Samples were analyzed in duplicate. Plasma concentrations were log-transformed and plotted against incubation time. Data were analyzed by simple linear regression, and the decomposition was given in terms of first-order rate constants.
To study the stability of the internal standard in TCA used for plasma protein precipitation, we studied aliquots of 100 g/L TCA containing 150 nmol/L isoriboflavin incubated for 1, 2, 4, 7, and 14 days. These aliquots were used for precipitating identical plasma samples. Riboflavin was added as a reference calibrator, and the samples were processed according to standard procedures.
We investigated the stability of endogenous riboflavin, FMN, and FAD in whole blood by incubating aliquots from one donor for 1, 2, and 5 h.
To study the stability of the riboflavin vitamers in plasma, we divided a sample containing endogenous vitamin concentrations into four batches, with 500 nmol/L riboflavin, FMN, or FAD added to three of the batches. Plasma aliquots were stored for 1, 4, 8, and 14 days before analysis.
We investigated the stability of the vitamers and the internal standard in TCA-treated plasma before neutralization by adding 500 nmol/L riboflavin, FMN, FAD, or isoriboflavin to the samples. The samples were then incubated for 1, 2, 4, 7, and 12 h before additional processing. The stability of the vitamers and the internal standard was also tested in neutralized TCA-treated plasma. Plasma samples to which 500 nmol/L riboflavin, FMN, FAD, or isoriboflavin had been added were analyzed after incubation for 1, 2, 4, and 7 days. In addition, the stability of the riboflavin vitamers and the internal standard was tested in the methanol-phosphate buffer obtained by solid-phase extraction using the same concentrations of riboflavin, FMN, FAD, or isoriboflavin and incubation for 1, 2, 4, and 7 days.
hemolysis
Hemolyzed EDTA blood was prepared by sonication (20 kHz for 5
s at 150 W). The hemoglobin concentration of the plasma was measured
(Technicon H2TM; Technicon Instruments), and the
percentage of hemolysis was calculated by comparing the hemoglobin
concentration of the plasma to the total hemoglobin concentration of
the blood. Sonicated and nonsonicated blood from the same donor was
mixed in different proportions to obtain different degrees of
hemolysis. Samples were incubated for 1, 2, and 5 h.
recovery and analytical variation
The recovery was determined by the addition of riboflavin (10 and
30 nmol/L), FMN (10 and 30 nmol/L), and FAD (50 and 100 nmol/L) to
plasma containing endogenous analyte. At each of the three
concentrations, 15 replicates were analyzed in one analytical
run.
To determine within-day precision, we used the data from the recovery experiments. Between-day precision was determined by assaying the same samples on 15 different days over a period of 3 weeks.
linearity and limit of detection
The linearity of the assay at vitamin concentrations greater than
the endogenous concentrations was determined by the addition of
increasing amounts of the vitamers, corresponding to plasma
concentrations between 0.3 and 1000 nmol/L greater than the endogenous
concentrations, to plasma containing endogenous concentrations of
riboflavin, FMN, and FAD.
To determine the limit of detection and assay linearity at concentrations less than the endogenous concentrations, we used photodegradation to prepare plasma devoid of riboflavin. Plasma proteins were precipitated with 100 g/L TCA without internal standard and then exposed to ultraviolet irradiation at 370 nm for 6 h; neutralizing solution and internal standard were then added. Plasma containing endogenous concentrations of the analytes was not irradiated, but otherwise was prepared in the same way and mixed with irradiated plasma in different proportions ranging from 0% to 100% to obtain different concentrations of the vitamers.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. The analytes eluted in the order riboflavin, lumiflavin, FMN,
and FAD and were baseline-resolved within 12 min and clearly separated
from the internal standard. The migration profile was stable over time,
and the within- and between-day relative standard deviations (RSD) of
the migration times were 1.5% and 2.0%, respectively.
|
sample processing and stability
Stability of the internal standard and the riboflavin
vitamers.
The stability of the internal standard was investigated
in 100 g/L TCA and at all stages of sample preparation. The stability
of riboflavin, FMN, and FAD was tested in EDTA whole blood, hemolyzed
blood, and EDTA plasma and at all stages of sample preparation. In all
experiments, samples were incubated in the dark at 23 °C.
The internal standard was stable for 14 days when dissolved in the 100 g/L TCA solution used for protein precipitation.
The vitamers were stable for at least 5 h in whole blood and for at least 1 h in blood with 1% hemolysis. This degree of hemolysis did not affect plasma flavin concentrations.
Plasma containing endogenous vitamin concentrations or supplemented with 500 nmol/L riboflavin, FMN, or FAD could be stored for 14 days without changes in the concentrations of riboflavin or FMN. In nonsupplemented plasma, FAD was stable for 14 days, whereas the added FAD degraded at a rate of ~0.03/day.
Riboflavin, FMN, and the internal standard were stable for 12 h in TCA-treated plasma before neutralization, whereas FAD was degraded at a rate of ~0.17/h under these conditions. All vitamers and the internal standard were stable for at least 2 days in neutralized TCA-treated plasma and in the methanol-phosphate buffer obtained by solid-phase extraction.
assay performance
Recovery and analytical variation.
The recoveries of
riboflavin, FMN, and FAD added to plasma at two concentrations were
90–103% (Table 1
). The within-day and between-day CVs were 4–9% and 6–12%,
respectively (Table 2
).
|
|
Linearity and lower limit of detection.
The linearity of the
assay was tested at concentrations greater than and less than the
endogenous vitamer concentrations (Fig. 2
), which were 14.0 nmol/L for riboflavin, 7.5 nmol/L for FMN,
and 85.5 nmol/L for FAD. The linearity was documented for
concentrations up to 200 nmol/L greater than the endogenous
concentrations for all vitamers. At concentrations lower than the
endogenous concentrations, linearity was observed down to
concentrations approaching the limit of detection, which was defined as
a signal-to-noise ratio >5:1.
|
application of the method
Vitamer concentrations in healthy adults.
Riboflavin, FMN, and
FAD were measured in plasma from 63 healthy volunteers: 35 men and 28
women. Thirty-six subjects were blood donors, and 27 subjects were
recruited from laboratory staff. The mean age was 42 years (range,
23–64 years).
The median concentration (5–95 percentiles) was 8.6 nmol/L (2.7–42.5
nmol/L) for riboflavin, 7.0 nmol/L (3.5–13.3 nmol/L) for FMN, and 57.9
nmol/L (44.5–78.1 nmol/L) for FAD (Fig. 3
). Accordingly, the interindividual variation was substantial
for riboflavin, whereas it was lower for FMN and, in particular, for
FAD. The vitamin concentrations were not related to age (P
>0.26, Spearman correlation) or gender (P >0.11,
t-test). There was a significant positive correlation
between the concentrations of riboflavin and FMN (r =
0.72; P <0.0001) and between FMN and FAD (r
= 0.37; P <0.003), whereas the correlation between
riboflavin and FAD (r = 0.13; P = 0.3)
was not significant.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
High sensitivity is obtained by exploiting the native fluorescence of
riboflavin and its derivatives, the excitation spectra of which overlap
the emission wavelength of 488 nm of the argon laser (21).
The lower sensitivity for FAD (Figs. 1
and 2
) is related to its lower
fluorescence yield compared with riboflavin and FMN (21). In
its present format, the method allows the detection of vitamin
concentrations far below the reference values obtained by us (Fig. 2
),
and therefore should be suitable for the assessment of riboflavin
status under deficiency states. Sensitivity enhancement might be
obtained by the use of a HeCd laser emitting at 442 nm, which is closer
to the excitation maximum of the vitamers. Furthermore, only ~25 nL
of the 420-µL eluate from the solid-phase columns is injected into
the capillary, and the mass sensitivity could be enhanced by
concentrating the eluate before the CE step.
A major portion of the riboflavin vitamers is protein bound(22)(23), and TCA treatment of plasma offers protein precipitation and probably extraction of the protein-bound vitamin fraction. The high ionic strength obtained by this procedure is incompatible with CE, but solid-phase extraction provides a means for sample desalting and for concentration of analytes before CE. We also evaluated organic solvents, e.g., acetonitrile, as protein precipitating agents, but this matrix prevented the retention of analytes on the reversed-phase columns.
Isoriboflavin is an isomeric form of riboflavin
(8-demethyl-6-methylriboflavin) and has been used previously as an
internal standard in HPLC assays (15). We did not detect
endogenous isoriboflavin in human plasma. Isoriboflavin is stable
during sample processing and co-elutes with the analytes during
solid-phase extraction. Inclusion of the internal standard corrects for
variable recovery during solid-phase extraction, differences in the
amount of sample injected into the capillary, and long-term changes in
laser output, and it reduces the within- and between-day CVs from
9–18% (data not shown) to 4–12% (Table 2
).
We investigated the stability of the vitamers and the internal standard
at different stages of sample processing because analyte degradation
would seriously impair assay stability. In neutralized TCA-treated
plasma, analytes were stable for at least 2 days at room temperature.
This may partly explain the high analytical recovery, which was close
to 100% (Table 1
), and acceptable within- and between-day CVs (Table 2
).
We also demonstrated that the riboflavin vitamers were stable for at least 5 h in whole blood and for several days in plasma at room temperature. This is in agreement with data published by Burch et al.(24), who found that the vitamers were stable in whole blood stored at 4 °C for 48 h. They also reported that endogenous FAD was stable in serum stored at room temperature for several hours, whereas 80% of FAD added to serum was hydrolyzed within 1 h. Others have found that FAD added to whole blood(25)(26) or plasma(16)(26) is highly unstable and is degraded within minutes. FMN added to whole blood was found to be decomposed within minutes by Nogami et al. (25), but more slowly by other investigators (26)(27); it appears to be relatively stable in plasma(16)(22)(26).
The apparent discrepancies between our stability data and those published by others (16)(22)(24)(25)(26)(27) can be attributed in part to different incubation temperatures or anticoagulants. Notably, we used EDTA, which has been shown to inhibit the activity of enzymes that hydrolyze FAD and FMN (28). The binding of the vitamers to proteins (22)(23) may protect these compounds from degradation, which explains the different stabilities of endogenous vs added vitamers in plasma.
Hemolysis may influence analyte concentrations in plasma, either by the release of flavins from blood cells or by the leakage of cellular enzymes that catalyze the conversion of the vitamers. We found that hemolysis up to 1% does not represent an analytical problem. This percentage of hemolysis is seldom encountered in the clinical setting, and it is higher than the routinely used percentage for interference testing of analytical methods (29).
Our data on analyte stability are the basis for prolonged, unattended sample processing and automated CE injection. In addition, the stability of these analytes allows the measurement of vitamins in clinical samples transported or stored for days as well as epidemiological studies based on frozen samples. We have analyzed plasma samples kept at -80 °C for 32 months. In these samples, the vitamer concentrations, including FAD, were similar to those observed in fresh samples (data not shown).
We found vitamin concentrations similar to those reported by others(11)(15)(30)(31). We also observed a low interindividual variation of FAD compared with variable amounts of riboflavin and FMN. Riboflavin, in particular, varied considerably between subjects (15)(31). This may indicate that the concentration of FAD is tightly regulated(24)(32)(33)(34), whereas plasma riboflavin may vary in response to recent vitamin intake (31). Thus, the different flavin species in plasma appear to be under separate regulatory or nutritional influence, and the measurement of all vitamers in plasma may be desirable. The measurement of all vitamers may also be important in the investigation of conditions under which the activities of the enzymes that catalyze the interconversion of the different vitamers are changed, e.g., in riboflavin deficiency(34)(35) or thyroid disease(36)(37).
The widely used erythrocyte glutathione reductase activity coefficient is an indirect measure of the FAD concentration in the erythrocytes (7)(8). It does not, however, determine the concentrations of riboflavin and FMN or the relative distribution of different flavin species. It is considered a sensitive and robust index of riboflavin deficiency, but is less suitable for the assessment of riboflavin status at high riboflavin intake(6). Furthermore, the erythrocyte glutathione reductase activity coefficient may give misleading results for certain conditions in which erythrocyte flavin metabolism is altered, e.g., glucose-6-phosphate dehydrogenase deficiency(6)(7) and ß-thalassemia (38).
In conclusion, we have constructed a robust CE method for the determination of riboflavin species in human plasma. The monitoring of native fluorescence with LIF detection affords sufficiently high sensitivity to measure vitamin concentrations far lower than the concentrations found in healthy subjects. The stability of the vitamers during sample processing and an almost complete recovery allows precise quantification of all three vitamers. The method thus may become a valuable tool to assess riboflavin status in humans.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  J.-Y. Hautem, C. Morel, R. Couderc, and F. Moussa Liquid chromatographic determination of b2 vitamers in human plasma and whole blood. Clin. Chem., May 1, 2006; 52(5): 907 - 908. [Full Text] [PDF]
|
|
|
|
|
|
O. Midttun, S. Hustad, E. Solheim, J. Schneede, and P. M. Ueland Multianalyte Quantification of Vitamin B6 and B2 Species in the Nanomolar Range in Human Plasma by Liquid Chromatography-Tandem Mass Spectrometry Clin. Chem., July 1, 2005; 51(7): 1206 - 1216. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Hustad, B. G Nedrebo, P. M. Ueland, J. Schneede, S. E. Vollset, A. Ulvik, and E. A Lien Phenotypic expression of the methylenetetrahydrofolate reductase 677C->T polymorphism and flavin cofactor availability in thyroid dysfunction Am. J. Clinical Nutrition, October 1, 2004; 80(4): 1050 - 1057. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Hustad, M. C. McKinley, H. McNulty, J. Schneede, J.J. Strain, J. M. Scott, and P. M. Ueland Riboflavin, Flavin Mononucleotide, and Flavin Adenine Dinucleotide in Human Plasma and Erythrocytes at Baseline and after Low-Dose Riboflavin Supplementation Clin. Chem., September 1, 2002; 48(9): 1571 - 1577. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Hustad, P. M. Ueland, S. E. Vollset, Y. Zhang, A. L. Bjorke-Monsen, and J. Schneede Riboflavin as a Determinant of Plasma Total Homocysteine: Effect Modification by the Methylenetetrahydrofolate Reductase C677T Polymorphism Clin. Chem., August 1, 2000; 46(8): 1065 - 1071. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
