Clinical Chemistry 51:6 1007–1013 (2005)
Automation and Analytical Techniques
Speciation and Quantification of Thiols by Reversed-Phase Chromatography Coupled with On-Line Chemical Vapor Generation and Atomic Fluorescence Spectrometric Detection: Method Validation and Preliminary Application for Glutathione Measurements in Human Whole Blood Emilia Bramanti,1* Cecilia Vecoli,2 Danilo Neglia,2 Maria Paola Pellegrini,2 Giorgio Raspi,1 and Renata Barsacchi2,3
Background: We developed a sensitive, specific method for the low–molecular-mass thiols cysteine, cysteinylglycine, glutathione, and homocysteine and validated the method for measurement of glutathione in blood. Methods: The technique was based on reversed-phase chromatography (RPC) coupled on line with cold vapor generation atomic fluorescence spectrometry (CVGAFS). Thiols were derivatized before introduction on the column by use of a p-hydroxymercuribenzoate (PHMB) mercurial probe and separated as thiol-PHMB complexes on a Vydac C4 column. Postcolumn on-line reaction of derivatized thiols with bromine allowed rapid conversion of the thiol-PHMB complexes to inorganic mercury with recovery of 100 (2)% of the sample. HgII was selectively detected by atomic fluorescence spectrometry in an Ar/H2 miniaturized flame after sodium borohydride reduction to Hg0.
1 Italian National Research Council, Istituto per i Processi Chimico-Fisici, Laboratory of Instrumental Analytical Chemistry, Pisa, Italy. 2 Italian National Research Council, Istituto di Fisiologia Clinica, Pisa, Italy. 3 Department of Physiology and Biochemistry, University of Pisa, Pisa, Italy. *Address correspondence to this author at: Italian National Research Council, Istituto per i Processi Chimico-Fisici, Laboratory of Instrumental Analytical Chemistry, Via G. Moruzzi 1, 56124 Pisa, Italy. Fax 39-050-315-2555; e-mail [email protected]
Received November 16, 2004; accepted March 29, 2005. Previously published online at DOI: 10.1373/clinchem.2004.045443
Results: The relationship between thiol-PHMB complex concentration and peak area (CVGAFS signal) was linear over the concentration range 0.01–1400 mol/L (injected). The detection limits were 1, 1, 0.6, and 0.8 nmol/L for cysteine, cysteinylglycine, homocysteine, and glutathione in the injected sample, respectively. The CVs for thiols were 1.5%–2.2% for calibrator solutions and 2.1% and 3.0% for real samples. The RPCCVGAFS method allowed speciation of glutathione (reduced and oxidized) in human whole blood from healthy donors and from the coronary sinus of patients with idiopathic dilated cardiomyopathy during and after chronotropic stress. Conclusion: The RPC-CVGAFS method could be used to measure reduced and oxidized glutathione in human whole blood as disease biomarkers. © 2005 American Association for Clinical Chemistry
Low–molecular-mass thiols, such as cysteine (Cys), cysteinylglycine (CysGly),4 reduced glutathione [␥-l-glutamyl-l-cysteinylglycine (GSH)], and homocysteine (Hcy), and their disulfides are critical cellular components
4 Nonstandard abbreviations: CysGly, cysteinylglycine; GSH, reduced glutathione (␥-l-glutamyl-l-cysteinylglycine); Hcy, homocysteine; PHMB, phydroxymercuribenzoate; DTT, dithiothreitol; GSNO, nitrosoglutathione; GSSG, glutathione disulfide; TFA, trifluoroacetic acid; RPC, reversed-phase chromatography; DCM, dilated cardiomyopathy; DAD, diode array detector; CVGAFS, cold vapor generation atomic fluorescence spectrometry; and LODc, concentration detection limit.
Bramanti et al.: Thiol Speciation and Quantification by RPC-CVGAFS
that play numerous important roles in metabolism and homeostasis. These compounds act as part of an antioxidant defense network, including radical quenching, and altered thiol concentrations in plasma have been linked to specific pathologic conditions (1 ). Glutathione is the principal nonprotein thiol compound, acting as a major bio-reducing agent (2 ). The intracellular concentration of GSH in mammalian cells is in the millimolar range (0.5–10 mmol/L), whereas micromolar concentrations are typically found in plasma (3 ). Cys is a critical substrate for protein synthesis and a rate-limiting precursor of GSH and taurine synthesis; with the other thiols (in particular GSH), it is involved in the transport/storage of NO (4 ). Hcy is a sulfur-containing amino acid that is formed from methionine, an essential amino acid derived from dietary protein. More than 80% of Hcy is found in plasma, mostly conjugated to proteins through disulfide bonding or as symmetrical disulfide homocystine, as mixed disulfide Hcy-Cys, or as a free thiol (⬍2%) (1, 5 ). Mildly increased plasma Hcy concentrations have been associated with an increased risk for cardiovascular and cerebrovascular diseases in men (6 ), and highly increased plasma and urine concentrations are a clinically relevant indicator of a rare, hereditary metabolic disorder called homocystinuria (7 ). CysGly is, after Cys, the most abundant thiol in plasma and is a product of enzymatic degradation of glutathione. Reduced, free oxidized, and protein-bound forms of Cys, CysGly, glutathione, and Hcy comprise the plasma redox thiol status (8 ). Although they play important roles in metabolism and homeostasis and their concentrations in blood (plasma and erythrocytes) could possibly serve as biomarkers of redox status, biological thiols and disulfides are present in blood in low concentrations, and blood is a complex matrix. There thus is a need to develop new rapid, sensitive, specific, and accurate methods for their determination. Several methods have been proposed for the measurement of thiols and disulfides in plasma, most of them based on derivatization procedures (1, 9 ). A review of methods for GSH and related thiols has been published (3 ). Here we present a new, abbreviated technique that has been optimized for the speciation and determination of the thiols Cys, CysGly, glutathione, and Hcy.
Materials and Methods chemicals and stock solutions p-Hydroxymercurybenzoate [4-(hydroxymercuric)benzoic acid, sodium salt; PHMB] was purchased from Sigma. We prepared a 0.01 mol/L stock solution of PHMB by dissolving the sodium salt in 0.01 mol/L NaOH to improve its solubility; we stored the solution at 4 °C and diluted it just before use. The precise PHMB concentrations in the solutions were determined from the absorbance at 232 nm (⑀232 ⫽ 1.69 ⫻ 104 cm⫺1 䡠 mol⫺1 䡠 L). l-Cys was purchased from Merck, and reduced GSH,
d,l-Hcy, dithiothreitol (DTT), nitrosoglutathione (GSNO), and CysGly were purchased from Sigma. Glutathione disulfide (GSSG) was purchased from Fluka. Perchloric acid and EDTA disodium salt were from J.T. Baker. We prepared a 0.2 mol/L (pH 9.0) buffer solution from NaH2PO4 monohydrate and anhydrous Na2HPO4 (BDH; Merck). Trifluoroacetic acid (TFA) and methanol for reversed-phase chromatography (RPC) were purchased from Carlo Erba. We prepared NaBH4 solutions more concentrated than 10 g/L by dissolving the solid reagent (Merck; minimum assay ⬎96%) in a solution of 3 g/L NaOH. The solutions were microfiltered through a 0.45 m membrane and stored in a refrigerator. We prepared dilute solutions of NaBH4 by appropriately diluting the stock solutions, the total NaOH concentration being kept at 3 g/L, unless specified. A 240 –260 g/kg hydrazine standard solution (cat. no. 53847) was purchased from Fluka Chemie, and the optimized concentration was added to a NaBH4 solution containing 3 g/L NaOH (see Table 1 in the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol51/ issue6/). We prepared the stock solution of Br⫺/BrO3⫺ by dissolving the solid reagents (Carlo Erba), keeping an ⬃5:1 molar ratio on the basis of stoichiometry of redox reaction. Addition of a moderate excess of Br⫺ guaranteed complete conversion of bromate to Br2. Water deionized with a Milli-Q system (Millipore) was used throughout.
sample collection and storage We obtained human blood (1.0 mL) from healthy volunteers by venipuncture of an antecubital vein with a butterfly needle and collected the blood into evacuated tubes containing EDTA; we obtained blood from cardiopathic patients by catheterization of the coronary sinus after obtaining informed consent. The protocol was approved by the internal scientific review board and ethics committee. In brief, 5 patients with left ventricular dysfunction and angiographically normal coronary arteries at routine cardiac catheterization, receiving the diagnosis of idiopathic dilated cardiomyopathy (DCM), underwent additional catheterization of the coronary sinus. Blood samples (1 mL) were collected from the coronary sinus by syringe from patients at rest, during atrial pacing (pulse, 110 and 130), and in the recovery phase (1, 5, 15, and 30 min) to assess markers of oxidative stress; the blood samples were immediately injected into evacuated tubes containing EDTA. Blood samples were obtained, after consent, from 5 control patients (with coronary artery disease and normal left ventricular function) by venipuncture of an antecubital vein with a butterfly needle and were collected in evacuated tubes containing EDTA. Samples were taken with the patients in a resting condition at 0, 5, 15, and 30
Clinical Chemistry 51, No. 6, 2005
min to assess the within- and between-person variability of GSH status over time. For analysis of reduced and oxidized glutathione in whole blood, we mixed 500 L of the collected blood with 500 L of 100 g/L trichloroacetic acid immediately after collection and centrifuged the acidified sample at 10 000 g for 15 min at 4 °C. The supernatant was divided in aliquots and stored at ⫺80 °C until use.
sample derivatization procedure Fresh thiol stock solutions (⬃2–5 g/L) were prepared daily by dissolving the powder in 1 mL/L TFA. For calibration experiments, thiols were derivatized by diluting the stock solution in 1 mL/L TFA containing a stoichiometric amount or a moderate excess of PHMB (25 °C). After a reaction time of 50 min at room temperature (20 ⫾ 1 °C), the solutions were injected into the RPC system. For each blood sample, we diluted 2 aliquots (50 L each) of the supernatant coming from the deproteinized whole blood in 200 L of 0.2 mol/L sodium phosphate buffer (pH 9.0) and 1 mmol/L EDTA. One aliquot was mixed with 750 L of 1 mL/L TFA containing 100 mol/L PHMB for measurement of reduced GSH. To the other aliquot, we added 0.5 mmol/L DTT and incubated it at 37 °C for 20 min to reduce the GSSG for measurement of total GSH (GSHtot). After 20 min, the reaction was terminated by addition of 750 L of 1 mL/L TFA containing an excess of PHMB to complex the GSH and the DTT. The GSSG concentration was calculated as the difference between the concentrations of GSHtot and GSH. To evaluate the optimum DTT concentration for complete conversion of GSSG to GSH, we investigated the concentration–response curve for DTT in the reaction mixture, varying the DTT concentration from 0.1 to 10 mmol/L. We also studied the kinetics of DTT reduction; the reaction appeared to be complete at 20 min and remained unchanged after 30 min of incubation at 37 °C (data not shown). In the adopted conditions, the mean (SD) reduction yield was 98 (2)%.
stability of thiol-PHMB complexes All of the thiol solutions, once complexed, were stable for the length of the working day if kept at room temperature and for 3 month (time tested) if stored at ⫺20 °C.
chromatographic instrumentation For RPC, we used an HPLC gradient pump (P4000; ThermoQuest) equipped with a mechanical degassing system (SC1000; ThermoQuest), a Rheodyne 7125 injector (Rheodyne), and a 35-L injection loop. A diode array detector (DAD; UV6000; ThermoQuest) equipped with a flow cell with a 5-cm pathlength was located at the end of the column, just before and in series with the cold vapor generation atomic fluorescence spectrometry (CVGAFS) detection system. This detection system (DAD-CVGAFS)
allowed simultaneous acquisition of ultraviolet/visible absorbances and mercury-specific chromatograms.
chromatographic conditions Separations were carried out with a Vydac C4 reversedphase column [150 ⫻ 4.6 mm (i.d.); silica particle size, 5 m; porosity, 300 Å; code 214TP5415]. All solutions were filtered through a 0.45 m cellulose acetate filter (Millipore). The pump flow rate was 1.0 mL/min. Samples were eluted with a 45-min linear gradient from 100% A (1 mL/L TFA solution) to 30% B (per liter, 950 mL of methanol and 50 mL of 1 mL/L TFA solution). The PHMB-DTT complex eluted at 28 min and did not interfere with the measurement of other thiols.
chemical vapor generation with afs detection A schematic diagram of the continuous flow (CF) mercury chemical vapor generator modified for on-line digestion of organic mercury in a miniaturized Ar/H2 flame is shown in Fig. 1 of the online Data Supplement, and a detailed description of the apparatus has been reported previously (10 ). Briefly, on-line digestion was performed at room temperature in a 30-cm–long reaction coil with bromine, generated in situ by KBr/KBrO3 in HCl medium. The subsequent reduction of HgII to Hg0 was performed in a knitted reduction coil with NaBH4 solution containing hydrazine to control the potential quenching effect of Br2 on mercury fluorescence (10 ). The use of short knitted coils guaranteed effective mixing of reactants and control of diffusion processes with resulting symmetric, well-shaped chromatographic peaks (11 ). Two independent peristaltic pumps were used for mercury vapor generation and for waste removal from the gas–liquid separator. A laboratory-made manifold was used for mixing reagents and for addition of the argon carrier gas to the reaction mixture. The manifold consisted of a series of independent mixing T-junctions (0.5 mm i.d.) drilled into a Teflon block. The gas–liquid separator was made from borosilicate glass and was 60 mm long with a 10 mm i.d. All reagent concentrations, reaction coil dimensions, and flow rates were optimized, as reported previously (10 ), and are summarized in Table 1 of the online Data Supplement. The mercury vapor coming from the gas–liquid separator was delivered into the atomizer, which was a miniature Ar/H2 diffusion flame supported on a simple quartz tube (4 mm i.d.). Despite the fact that mercury vapor is in its atomic state and does not require the atomization step, we found that the Ar/H2 diffusion flame was necessary to remove interferences in the atomization/detection step generated by volatile species arising from the complex matrix during the postcolumn chemical reactions. We used a laboratory-assembled nondispersive atomic fluorescence detector (12 ) equipped with an EDL2 System (Perkin-Elmer) and an electrodeless discharge mercury lamp (EDL).
Bramanti et al.: Thiol Speciation and Quantification by RPC-CVGAFS
The output data from the lock-in amplifier were collected at 1 Hz with a personal computer (600 MHz Pentium®; Intel Corporation) equipped with a data acquisition card (PCL 718H; Advantech) and its acquisition software (Genie 2.12; Advantech). The ultraviolet data were averaged down to 5 points/s before saving; a 5-point boxcar averaging was applied to the atomic fluorescence signal data. Raw data were saved and processed by Origin 6.1 Professional (OriginLab Corporation).
Results and Discussion speciation and determination of PHMB complexes in solutions by RPC-CVGAFS Shown in Fig. 1 are RPC-CVGAFS chromatograms for Cys-, CysGly-, Hcy-, and GSH-PHMB complexes (6.5, 9.7, 6.5, and 8.0 mol/L in the sample injected, respectively), obtained with the optimized conditions. We constructed calibration curves for thiol-PHMB complexes (injected concentration, 0.01–1400 mol/L) by injecting calibrators in duplicate as described in the Materials and Methods. Calibration curves were linear over 3 orders of magnitude, the dynamic linear range spanning injected concentrations of 0.01–1400 mol/L. The fitting results of the calibration curves performed under the optimized conditions (Table 1) were obtained by plotting the areas of RPC-CVGAFS chromatogram peaks as a function of thiol-PHMB complex concentration injected. The retention times and concentration detection limits (LODc) are shown. The LODc values, calculated from the calibration curves and based on the concentrations giving a signal 3 times the baseline noise, were 1, 1, 0.6, and 0.8 nmol/L for Cys, CysGly, Hcy, and GSH, respectively.
Fig. 1. RPC-CVGAFS chromatographic profiles of Cys, CysGly, Hcy, and GSH complexed with PHMB. The concentrations were 6.5, 9.7, 6.5, and 8.0 mol/L, respectively, in the sample injected. AF, atomic fluorescence.
Table 1. Results of calibration experiments for thiol-PHMB complexes measured by the RPC-CVGAFS system. Sample
Retention time, min
Mean (SD) slope, V 䡠 min 䡠 molⴚ1 䡠 L
Cys-PHMB CysGly-PHMB Hcy-PHMB GSH-PHMB
8.53 9.29 14.07 17.59
0.0126 (0.00009) 0.0110 (0.00006) 0.0238 (0.00014) 0.0179 (0.00013)
0.9997 0.9998 0.9998 0.9997
10 11 9 11
1 1 0.6 0.8
The complexation ratio between the mentioned thiols and PHMB has been demonstrated to be 1:1; therefore, the slopes for the 4 thiol-PHMB complexes should be comparable. The different sensitivities shown in Table 1 are attributable to the different yields for on-line oxidation of these complexes by bromine in the oxidation coil (see Fig. 1 in the online Data Supplement). This could depend on the different stabilities of the thiol-PHMB complexes, which were attributable to the different pKa values for the ⫺SH groups of the thiols (13 ).
GSH and GSSG measurements in whole blood We applied the RPC-CVGAFS method to the measurement of GSH and GSSG in whole peripheral blood from 5 healthy donors. The mean (SD) values for GSH and GSSG concentrations were 539.4 (70.9) and 90.4 (52.7) mol/L, respectively, with the percentage of GSSG/GSHtot ranging between 1% and 21%. The within-day reproducibility of the assay for GSH was determined with 15 underivatized aliquots prepared independently from the whole-blood sample, and the between-day reproducibility was calculated from the analysis of 10 different underivatized aliquots of the supernatant from the whole-blood sample, prepared as described in the Materials and Methods and stored at ⫺80 °C, which was derivatized fresh each day for 10 days over a period of 1 month. The within-day CVs were 1.9% for GSH and 3.0% for GSSG. The between-day CVs, determined by assaying on 10 different days, were 2.8% for GSH and 4.0% for GSSG. We evaluated recovery by adding known concentrations of GSH or GSSG (400 mol/L) to whole blood from a healthy donor (N3). The concentrations in samples with added glutathione were measured in 5 replicates. The mean (SD) recoveries were 103 (5)% for GSH and 98 (6)% for GSSG. The RPC-DAD-CVGAFS chromatogram of a wholeblood sample (endogenous GSH, 23 mol/L), derivatized with PHMB and enriched with Cys-PHMB (19 mol/L in the sample injected), CysGly-PHMB (47 mol/L), and Hcy-PHMB (19 mol/L) is shown in Fig. 2. The recovery of the added analytes, calculated on the basis of peak areas and calibration curves in 5 enriched samples, was 98 (3)%. The black line in Fig. 2 corresponds to the mercury-specific chromatographic trace; the gray line indicates the ultraviolet absorbance chromatogram at 254 nm; comparison of the traces reveals the better response
Clinical Chemistry 51, No. 6, 2005
whether the eventual presence of GSNO in the sample affects the measurement of reduced GSH or of GSSG. For this purpose, we added a known concentration of GSNO (250 mol/L) to whole blood from healthy donors and treated the samples according to the procedure described in the Materials and Methods. We determined the recovery of GSNO in both nonreduced samples and in samples treated with DTT. The concentrations in samples with added GSNO were determined in 3 replicates. GSNO was not detected in the nonreduced sample, but was recovered in the sample treated with DTT [mean recovery 98 (2)%], indicating that the PHMB does not react with GSNO unless reduced by DTT. Thus, GSHtot can be expressed as:
Fig. 2. RPC-DAD-CVGAFS chromatographic profiles of a whole-blood sample derivatized with PHMB according to the procedure described in the Materials and Methods. The sample was enriched with Cys-PHMB (19 mol/L in the sample injected), CysGly-PHMB (47 mol/L), and Hcy-PHMB (19 mol/L). The GSH (23 mol/L in the sample injected) is endogenous. The black line corresponds to the mercuryspecific chromatographic trace (CVGAFS detection); the gray line indicates the ultraviolet (UV) absorbance at 254 nm.
[GSH]tot ⫽ [GSHred] ⫹ [GSHox]
[GSHox] ⫽ 2 [GSSG] ⫹ [GSNO]
When the GSNO concentration is not known, we suggest that the oxidized glutathione (GSHox) be expressed with more confidence as the difference from Eq. 1: [GSHox] ⫽ [GSH]tot ⫺ [GSHred]
of the CVGAFS detector compared with a conventional ultraviolet detector. In the specific case of detection of GSH in whole blood, the sensitivity of the RPC-CVGAFS method is much higher that that required. Indeed, the total GSH concentration in whole blood ranges between 0.5 and 0.7 mmol/L, most of which originates from the GSH inside erythrocytes with ⬃0.2% from plasma. For this reason, we diluted our samples 40-fold before injection. Dilution of complex biological matrices is recommended to avoid or control for any matrix effects. For measurement of the other thiols, greater sensitivity is required because Cys, CysGly, and Hcy are present in plasma and cells in the micromolar range.
behavior of GSNO in RPC-CVGAFS analysis It is well known that the incorporation of NO into S-nitroso derivatives of thiols, such as Cys and, in particular, GSH, may serve to maintain over time the biological activity of NO (14 ). Thus, the part of GSH in blood present as GSNO may have biological relevance. Recent studies have indicated that the concentrations of S-nitroso derivatives of thiols in plasma from healthy individuals are lower than reported previously and in the range of 30 –120 nmol/L (15, 16 ). Although in physiologic conditions GSNO represents a small part of total glutathione (GSHtot ⫽ GSH ⫹ GSNO ⫹ 2 GSSG), it should not be overlooked in thiol speciation studies because of the potential role of GSNO in many pathologic situations (17 ). Because the RPC-CVGAFS method is based on the interaction of a mercurial probe with GSH, it was important to investigate the behavior of GSNO to understand
GSH and GSHox in patients with idiopathic DCM during chronotropic stress We applied the RPC-CVGAFS method to the determination of reduced GSH and GSHox in blood samples from patients with DCM during chronotropic stress (18 ). The protocol included sampling at rest (time points B1 and B2 in Fig. 3), after 3 min of pacing at a pulse rate of 110 (time point P110), after 3 min of pacing at a pulse rate of 130 (time point P130), and during recovery (time points R1, R5, R15, and R30, after 1, 5, 15, and 30 min, respectively). We also evaluated the GSHox/GSHtot ratio in the peripheral blood of 5 control patients (with coronary artery disease and normal ventricular function) at 0, 5, 15, and 30 min to assess the within- and between-person variability of GSH status over a time period analogous to the chronotropic stress test. The trends of GSH and GSHox concentrations in the 5 DCM patients (Pt1, Pt2, Pt3, Pt4, and Pt5, respectively) under the different study conditions are shown in Fig. 2 of the online Data Supplement, whereas Fig. 3 shows the trends of the GSHox/GSHtot ratio in the same patients (panels A and B) and in the 5 controls (panel C). We observed no significant variations in GSH status in the 5 controls examined, with the GSHtot concentration ranging between 408 and 724 mol/L in the whole-blood sample, and the GSHox/GSHtot ratio ranging between 5% and 28%, without notable changes over time. In the DCM patients, GSH and GSHox concentrations and the GSHox/GSHtot ratio at rest (samples B1 and B2) were in the same range as for the controls, the GSHox/ GSHtot ratio ranging between 11% and 28%. However, in all 5 DCM patients, the GSHox/GSHtot ratio in the coro-
Bramanti et al.: Thiol Speciation and Quantification by RPC-CVGAFS
nary sinus blood tended to increase during chronotropic stress. In 1 of the 5 patients (Pt4 in Fig. 3B), the increase in the ratio was delayed and the recovery was unappreciable during the test time. In patient 5 (Fig. 3B, Pt5) the study was prematurely interrupted at 15 min of recovery because of displacement of the coronary sinus catheter and inability to sample blood. In the other 3 patients (Fig. 3A), the GSHox/GSHtot ratio tended to decrease during recovery. These data indicate that reduction of GSHox, de novo synthesis in erythrocytes of GSH, or GSH secretion by myocardium tissues or endothelial cells may have occurred in the recovery phase to counteract a putative oxidative stress induced by pacing in the myocardium. Emotional stress also affects the GSHox/GSHtot ratio. Of the 5 cases examined in the second sampling before the chronotropic stress (B2), the GSHox/GSHtot ratio was not altered in patients 1 and 4, but it reached 30% in patient 3, 55% in patient 5, and 70% in patient 2. No data have been reported in the literature to which our results can be compared, mainly because no studies measuring acute changes of markers of oxidative stress in patients with DCM have been performed. However, recent data suggest that in these patients, myocardial ischemia may occur during stress as a consequence of coronary microvascular dysfunction (19, 20 ). Accordingly, changes in markers of oxidative stress in DCM patients may reflect myocardial ischemia. It is known that GSH improves the recovery of myocardial mechanical function after ischemia–reperfusion in rat heart, an effect that may be related to the detoxification of peroxynitrite by GSH and the stimulation of soluble guanylate cyclase (21 ). In a dog model, it has been observed that, parallel to increased lipid peroxidation, reduced GSH gradually decreased in coronary sinus blood sampling during early reperfusion after myocardial ischemia (22 ).
comparison of RPC-CVGAFS with other methods
Fig. 3. Effect of chronotropic stress on the redox state of glutathione in whole blood from patients with DCM during chronotropic stress and in control patients. (A), GSHox/GSHtot ratio in patients 1, 2, and 3 (Pt1, Pt2, and Pt3); (B), GSHox/GSHtot ratio in patients 4 and 5 (Pt4 and Pt5). (C), GSHox/GSHtot ratio in 5 control patients with coronary artery disease and normal left ventricular function at 0, 5, 15, and 30 min.
With the described analytical conditions we could analyze 4 samples/h (run-to-run time of 15 min, including column equilibration). Cys and CysGly are not baseline separated (tR ⫽ 4.88 and 5.03 min, respectively), and Hcy elutes at 7.21 min. The cost of the mercurial probe PHMB is low, and PHMB-derivatized samples are stable at pH 2– 8.5 at ⫺20 °C. The proposed method, as well as all of the methods that use the specific derivatization of ⫺SH groups, is not capable of monitoring disulfides directly, requiring a reduction step for detection of oxidized thiols. A comparison of the LODc values for published methods for thiol speciation and determination with the LODc for the RPC-CVGAFS method is shown in Table 2 of the online Data Supplement. Detection limits for the RPCCVGAFS method are comparable to those obtained by liquid chromatography with electrochemical detection and enzymatic assays and are approximately 2 orders of magnitude higher than the LOD for capillary electrophoresis with laser-induced fluorescence detection. RPC-
Clinical Chemistry 51, No. 6, 2005
CVGAFS is rather inexpensive compared with capillary electrophoresis with laser-induced fluorescence detection and liquid chromatography–mass spectrometry, although liquid chromatography–mass spectrometry offers good sensitivity and unique selectivity. Electrochemical detection has the advantage of detecting both disulfides and reduced thiols, without requiring a derivatization step. However, most of the liquid chromatography detection methods for thiols require the use of mercury electrodes (23 ) rather than the more common carbon electrode (24 ) because thiols have a high overpotential for electrooxidation at ordinary carbon electrodes. In addition, Au/Hg electrodes (chemically modified or not) are characterized by long-term instability during both experimental operation and storage, and at high oxidation potentials, many compounds (e.g., oxygen) may cause interference (24, 25 ). Expensive reagents are necessary for immunoassaybased methods, although the labor costs are lower and the throughput is higher. However, simultaneous determination of thiols is not possible with immunoassays.
This work was supported financially by CNR, Istituto di Fisiologia Clinica, Pisa, Italy. We thank ThermoQuest for providing part of the instrumentation, and M. Cempini, R. Spiniello, and M.C. Mascherpa for technical support.
References 1. Mansoor MA, Svardal AM, Ueland PM. Determination of the in vivo redox status of cysteine, cysteinylglycine, homocysteine, and glutathione in plasma. Anal Biochem 1992;200:218 –29. 2. Singh RJ. Glutathione: a marker and antioxidant for aging. J Lab Clin Med 2002;140:380 –1. 3. Pastore A, Federici G, Bertini E, Piemonte F. Analysis of glutathione: implication in redox and detoxification. Clin Chim Acta 2003;333:19 –39. 4. Giustarini D, Milzani A, Colombo R, Dalle-Donne I, Rossi R. Nitric oxide, S-nitrosothiols and hemoglobin: is methodology the key? Trends Pharmacol Sci 2004;25:311– 6. 5. Zhloba AA, Blashko EL. Liquid chromatographic determination of total homocysteine in blood plasma with photometric detection. J Chromatogr B 2004;800:275– 80. 6. Jacobsen DW. Homocysteine and vitamins in cardiovascular disease. Clin Chem 1998;44:1833– 43. 7. Mudd H, Levy HL, Skorby F. Disorders of transulfuration. In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds. The metabolic and molecular bases of inherited disease, 7th ed. New York: McGrawHill, 1995:1279 –327. 8. Ueland PM, Mansoor MA, Guttormsen AB, Muller F, Aukrust P, Refsum H, et al. The Hordaland Homocysteine Study: the opposite tails odds ratios reveal differential effects of gender and intake of vitamin supplements at high and low plasma total homocysteine concentrations. J Nutr 1996;126:S1244 – 8. 9. Bald E, Chwatko G, Gowacki R, Kusmierek K. Analysis of plasma
thiols by high-performance liquid chromatography with ultraviolet detection. J Chromatogr A 2004;1032:109 –15. Bramanti E, Lomonte C, Galli A, Onor M, Zamboni R, Raspi G, et al. Characterization of denatured metallothioneins by reversed phase coupled with on line chemical vapour generation and atomic fluorescence spectrometric detection (RP-CVGAFS). J Chromatogr A 2004;1054:285–91. Kolthoff IM. Theoretical aspect of FIA. In: Ruzicka J, Hansen EH, eds. Flow injection analysis. New York: John Wiley & Sons, 1988:87–122. D’Ulivo A, Bramanti E, Lampugnani L, Zamboni R. Improving the analytical performance of hydride generation non-dispersive atomic fluorescence spectrometry. Combined effect of additives and optical filters. Spectrochim Acta B 2001;56:1893–907. Bramanti E, D’Ulivo A, Lampugnani L, Raspi G, Zamboni R. Cold vapour atomic fluorescence studies on the behaviour of Hg(II) and Hg(II)–thiol complexes. An alternative route for characterization of ⫺SH binding groups. J Anal At Spectrom 1999;14:179 – 85. Wolzt M, McAllister RJ, Davis D, Feelisch M, Moncada S, Vallance P, et al. Biochemical characterization of S-nitrosohemoglobin. J Biol Chem 1999;274:28983–90. Marley R, Feelisch M, Holt S, Moore K. A chemiluminescensebased assay for S-nitroso-albumin and other plasma S-nitrosothiols. Free Radic Res 2000;32:1–9. Goldman RK, Vlessis AA, Trunkey DD. Nitrosothiol quantification in human plasma. Anal Biochem 1998;259:98 –103. Ignarro LJ, Napoli C, Loscalzo J. Nitric oxide donors and cardiovascular agents modulating the bioactivity of nitric oxide. An overview. Circ Res 2002;90:21– 8. Marini C, Eicke J-S, Salvadori P, Sambuceti G, L’Abbate A, Neglia D. Chronotropic stress enhances the effects of asynchronous left ventricular activation on myocardial perfusion in patients with idiopathic dilated cardiomyopathy and left bundle branch block [Abstract]. J Am Coll Cardiol 2004;43:A231. Neglia D, Michelassi C, Trivieri MG, Sambuceti G, Giorgetti A, Pratali L, et al. Prognostic role of myocardial blood flow impairment in idiopathic left ventricular dysfunction, Circulation 2002; 105:186 –93. Van den Heuvel AFM, Van Veldhuisen DJ, Van der Wall EE. Regional myocardial blood flow reserve impairment and metabolic changes suggesting myocardial ischemia in patients with idiopathic dilated cardiomyopathy. J Am Coll Cardiol 2000;35:19 –28. Cheung PY, Wang W, Schulz R. Glutathione protects against myocardial ischemia–reperfusion injury by detoxifying peroxynitrite. J Mol Cell Cardiol 2000;32:1669 –78. Lantos J, Temes G, Go¨bo¨lo¨s L, Jaberansari MT, Ro ˜th E. Is peripheral blood a reliable indicator of acute oxidative stress following heart ischemia and reperfusion? Med Sci Monit 2001; 7:1166 –70. Shea D, MacCrehan WA. Determination of hydrophilic thiols in sediment porewater using ion-pair liquid chromatography coupled to electrochemical detection. Anal Chem 1988;60:1449 –54. Imai Y, Ito S, Fujita K. Determination of natural thiols by liquid chromatography after derivatization with 3,5-di-tert.-butyl-1,2-benzoquinone. J Chromatogr 1987;420:404 –10. Carro-Ciampi G, Hunt PG, Turner CJ, Wells PG. A high-performance liquid chromatographic assay for reduced and oxidised glutathione in embryonic, neonatal, and adult tissues using a porous graphite electrochemical detector. J Pharmacol Methods 1988; 19:75– 83.