309
Clinica Chimica Acta, 59 (1975) 309-312 @ Elsevier Scientific Publishing Company,
Amsterdam
-
Printed
in The Netherlands
CGA 6918
ESTIMATION OF SERUM URIC ACID BY HIGH PERFORMANCE CHROMATOGRAPHY WITH ELECTROCHEMICAL DETECTION
LAWRENCE
A. PACHLA
Department (Received
and PETER
T. KISSINGER
of Chemistry, Michigan State University, East Lansing, Mich. 48824 September
LIQUID
(U.S.A.)
13, 19 74)
Summary A precise method for serum uric acid is described based on direct electrochemical oxidation in the eluate from high performance liquid chromatography. The detection limit for uric acid was found to be approximately 1 pg and accurate measurements were possible at the 100 pg level. Detailed procedures are outlined for 0.5 ml and 25 ~1 serum samples, the relative standard deviations being 1.7% and 1.8%, respectively, for a 6.2 mg/dl serum pool. Samples were normally deproteinized, however, the analysis can be carried out by injection of 2 ~1 of serum diluted lo-fold with distilled water. The proposed micro method is highly selective and eliminates the need for the enzyme preparation and/or nonspecific calorimetric reagents in common use.
Introduction There continues to be considerable interest in improved assays for uric acid in serum [l] . Many variations on the classical calorimetric methods using redox indicators [2,3] and the more modern differential methods using the enzyme uricase [4,5] have been recently reported. Several novel electrochemical [6,7] and chromatographic [8,9] approaches have also been devised, however, these have not received wide attention. We have recently introduced an analyzer for clinical measurements based on the combination of high performance liquid chromatography with thin-layer electrochemistry [lo] . This combination (LCEC) can afford extreme sensitivity and excellent selectivity for monitoring electroactive organic molecules. For example, we have found this technique to be particularly useful for assay of urinary catecholamines in that many of the problems associated with fluorescence or gas chromatography can be avoided. Preliminary results from the use of LCEC to determine uric acid and ascorbic acid in urine have been reported [lo]. In the present paper procedures for serum uric acid are described which
310
offer several important advantages over previous assays, particularly for pediatric samples. The principles of the method are simple. A rapid anion exchange separation is followed by a selective flow through electrochemical cell. The electrooxidation reaction which occurs in the detector presumably involves a loss of 2 e- and 2 H’ to form a bis-imine which subsequently hydrolyzes primarily to alloxan and urea. Dryhurst has discussed the electrochemistry of uric acid and other biologically important purines at length [ 111. Materials
and Methods
R eagen ts 1. Uric acid stock solution, 50 mg/dl: Dissolve 250 mg of 99% uric acid in about 250 ml water containing 300 mg of lithium carbonate by warming to about 60°C. Allow the solution to cool, then quantitatively transfer into a 500-ml volumetric flask and dilute to volume. 2. Uric acid working standard, 0.5 mg/dl. 3. Sulfuric acid, 0.033 M. 4. Sodium tungstate, 5%. 5. Mobile phase, pH 4.6,0.05 M acetate buffer.
Apparatus All liquid chromatographic data were obtained from an instrument previously described in detail [lo]. A Zipax (E.I. du Pont de Nemours and Co., No. 820960005) pellicular strong anion exchange packing was used in a 30 cm by 2 mm i.d. glass column at a flow rate of 0.30 ml/min. The detector potential was set at +0.80 V versus an Ag/AgCl reference electrode. For some samples a precolumn (1 cm long X 1 mm i.d.) packed with the stationary phase was used.
Liquid chromatography
procedure
To a 12 ml glass centrifuge tube add 4.0 ml of 0.033 M sulfuric acid and 0.5 ml of 5% sodium tungstate. While mixing on a vortex add 0.5 ml of serum and continue mixing for 40 seconds. Centrifuge the mixture to precipitate proteins and decant the clear supematant into a sample vial. Inject two microliters of sample onto the chromatographic column and record the peak area by electronic integration. The serum level in mg/dl is calculated by comparison with the integrated response to a 2 ~1 injection of the working standard. In the present work, standards were injected following every sixth sample although it is likely that much less frequent standardization would be necessary under routine operation of the LCEC instrument. When the supply of serum is limited, as in pediatric cases, the procedure is modified as follows to accommodate a 25 1.11sample. To a 1 ml centrifuge tube add 200 ~1 of 0.033 M sulfuric acid and 25 1.11of 5% sodium tungstate. While mixing on a vortex mixer add 25 ~1 of serum and continue mixing for about 3 minutes. Centrifuge and remove 2 ,ul of the clear supematant for direct injection into the instrument. A linear response of peak area to serum uric acid concentration was obtained up to at least 25 mg/dl for both procedures. The instrument sensitivity is
311
such that as little as 100 pg of uric acid can be determined quantitatively. In principle, it is therefore possible to reduce the volume of serum required for the protein precipitation step in the above micro procedure to about 0.25 ~1. The protein precipitation is not difficult to carry out on this scale; however, it is also practical to dilute the serum lo-fold with distilled water and inject it directly onto the column. We have devised a short precolumn to prevent clogging the narrow analytical column with protein. The precolumn can be easily replaced every few days, thus making deproteinization unnecessary. This approach works well and for routine situations the improvement in convenience is a decided advantage over previous methods. At present we are able to analyze 20 samples per hour on the chromatograph. The instrument is interfaced to a Digital Equipment Corporation S/E computer which automatically acquires and processes the data. The program was modified according to the recommendations of Chilcote and Mrochek [ 121 and is written in SABRE and FORTRAN 2. The details of the interface and a listing of the program are available on request.
AutoAnalyzer
procedures
Standard single channel AutoAnalyzer methods were used to corroborate the present results. The procedures used at both sites were standard colorimetric methods based on the reduction of phosphotungstic acid to tungsten blue. At one location (Upjohn Company) the classical chemistry of Brown [ 131 was used incorporating sodium cyanide in the dialyzed serum presumably to intensify the color and minimize the possibility of turbidity. This variation was not used at the other location (Erie County Laboratory) [14]. Results
and Discussion
The liquid chromatography of serum uric acid is essentially the same as that reported earlier for urine samples [lo]. The need for selectivity is more severe in the urine assay due to the higher concentrations of oxidizable metabolites. Combining the resolution of high performance liquid chromatography with the selectivity of electrochemistry results in great overall specificity for uric acid. The interferences normally implicated in the phosphotungstic acid method (i.e. ascorbic acid, xanthines, cysteine, glutathione, L-DOPA, and gentisic acid) do not contribute significantly to the present method even when present at lOO-fold the normal physiological concentrations. Corroboration of the specificity for uric acid was also carried out by enzymatic (uricase) removal of uric acid from a serum pool [15] prior to analysis by LCEC. No detectable signal was obtained for samples pretreated in this way. The precision obtained for repeated analysis of 0.5 ml and 25 ~1 aliquots of a 6.2 mg/dl serum pool was 1.7% and 1.8%, respectively. A correlation coefficient greater than 0.99 was obtained for both collaborative AutoAnalyzer studies using the phosphotungstic acid method. Fig. 1 illustrates the data for serum samples assayed by both LCEC and the calorimetric method which includes cyanide. Liquid chromatography has not yet been recognized as a routine tool for clinical analysis and it would therefore be difficult for many laboratories to
Fig.
1.
Correlation
between
liquid
chromatography
(y)
and
AutoAnalyzer
(I’)
procedures
for
104
serum
samples.
adopt the present method. Nevertheless, this situation is likely to change in the near future. The manifest advantages of LCEC may well prove it to be an attractive alternative to calorimetry in some cases. Acknowledgement The authors are indebted to Dr Terry J. Gilbertson (Upjohn Company, Kalamazoo, Mich. U.S.A.) and Dr David C. Wenke (Erie County Laboratory, Buffalo, N.Y.) for collaboration on the AutoAnalyzer methods. This investigation was supported by a grant from the National Science Foundation (GP42452X). References 1
R.
2
M.G.
Wells,
Clin.
Martinek.
J. Am.
Med.
Chim.
3
L.G.
Mm-in,
Clin.
4
N. Gochman
and
5
S. M&es.
C. Thompson
6
R.J.
and
Troy
7
G. Park,
8
P.A.
9
W.N.
R.N.
Kelly
Chem.. J.M.
W.C. Clin.
and
Kissinger.
and
Purdy, and
(1974) R.W.
W.R. 16
L.J.
Felice,
11
G. Dryhurst,
12
D.D.
13
H.J.
Brown,
J. Biol.
14
R.J.
Henry,
C. Sobel
15
D.G.
Remp,
Forts&r. and
Stand.
J.E.
Chim.
and
Forsch..
J. Kim,
Methods
Acta, Anal. Chem.,
Riggin.
158 Clin.
17
Clin.
(1971)
1154
Chem.,
27
20
(1970)
L&t..
(1974)
790
401
5 (1972)
887
191
Clin.
Mrochek,
Chem.,
Chem.,
White,
Chem.
233
379
Roach,
(1970)
R.M.
(1970)
51
Clin.
J. Wyngaarden.
P.T.
32 (1968)
Clin.
Chem.,
10
Chilcote
20
22
Schmitz.
Adams
Simkin,
Technol.,
Acta.
16
L.A. 34
(1972)
Clin.
Chem.,
(1945)
601
Am.
J. Clin.
Chem..
(1970)
Pachla 17
D.C.
Wenke,
Clin.
Chem.,
47 (1971)
Path&
6 (1970)
707
and
751 28
1
(1957)
152
and
645
20
(1974)
992
Clinica Chimica Acta, 59 (1975) 309-312 @ Elsevier Scientific Publishing Company,
Amsterdam
-
Printed
in The Netherlands
CGA 6918
ESTIMATION OF SERUM URIC ACID BY HIGH PERFORMANCE CHROMATOGRAPHY WITH ELECTROCHEMICAL DETECTION
LAWRENCE
A. PACHLA
Department (Received
and PETER
T. KISSINGER
of Chemistry, Michigan State University, East Lansing, Mich. 48824 September
LIQUID
(U.S.A.)
13, 19 74)
Summary A precise method for serum uric acid is described based on direct electrochemical oxidation in the eluate from high performance liquid chromatography. The detection limit for uric acid was found to be approximately 1 pg and accurate measurements were possible at the 100 pg level. Detailed procedures are outlined for 0.5 ml and 25 ~1 serum samples, the relative standard deviations being 1.7% and 1.8%, respectively, for a 6.2 mg/dl serum pool. Samples were normally deproteinized, however, the analysis can be carried out by injection of 2 ~1 of serum diluted lo-fold with distilled water. The proposed micro method is highly selective and eliminates the need for the enzyme preparation and/or nonspecific calorimetric reagents in common use.
Introduction There continues to be considerable interest in improved assays for uric acid in serum [l] . Many variations on the classical calorimetric methods using redox indicators [2,3] and the more modern differential methods using the enzyme uricase [4,5] have been recently reported. Several novel electrochemical [6,7] and chromatographic [8,9] approaches have also been devised, however, these have not received wide attention. We have recently introduced an analyzer for clinical measurements based on the combination of high performance liquid chromatography with thin-layer electrochemistry [lo] . This combination (LCEC) can afford extreme sensitivity and excellent selectivity for monitoring electroactive organic molecules. For example, we have found this technique to be particularly useful for assay of urinary catecholamines in that many of the problems associated with fluorescence or gas chromatography can be avoided. Preliminary results from the use of LCEC to determine uric acid and ascorbic acid in urine have been reported [lo]. In the present paper procedures for serum uric acid are described which
310
offer several important advantages over previous assays, particularly for pediatric samples. The principles of the method are simple. A rapid anion exchange separation is followed by a selective flow through electrochemical cell. The electrooxidation reaction which occurs in the detector presumably involves a loss of 2 e- and 2 H’ to form a bis-imine which subsequently hydrolyzes primarily to alloxan and urea. Dryhurst has discussed the electrochemistry of uric acid and other biologically important purines at length [ 111. Materials
and Methods
R eagen ts 1. Uric acid stock solution, 50 mg/dl: Dissolve 250 mg of 99% uric acid in about 250 ml water containing 300 mg of lithium carbonate by warming to about 60°C. Allow the solution to cool, then quantitatively transfer into a 500-ml volumetric flask and dilute to volume. 2. Uric acid working standard, 0.5 mg/dl. 3. Sulfuric acid, 0.033 M. 4. Sodium tungstate, 5%. 5. Mobile phase, pH 4.6,0.05 M acetate buffer.
Apparatus All liquid chromatographic data were obtained from an instrument previously described in detail [lo]. A Zipax (E.I. du Pont de Nemours and Co., No. 820960005) pellicular strong anion exchange packing was used in a 30 cm by 2 mm i.d. glass column at a flow rate of 0.30 ml/min. The detector potential was set at +0.80 V versus an Ag/AgCl reference electrode. For some samples a precolumn (1 cm long X 1 mm i.d.) packed with the stationary phase was used.
Liquid chromatography
procedure
To a 12 ml glass centrifuge tube add 4.0 ml of 0.033 M sulfuric acid and 0.5 ml of 5% sodium tungstate. While mixing on a vortex add 0.5 ml of serum and continue mixing for 40 seconds. Centrifuge the mixture to precipitate proteins and decant the clear supematant into a sample vial. Inject two microliters of sample onto the chromatographic column and record the peak area by electronic integration. The serum level in mg/dl is calculated by comparison with the integrated response to a 2 ~1 injection of the working standard. In the present work, standards were injected following every sixth sample although it is likely that much less frequent standardization would be necessary under routine operation of the LCEC instrument. When the supply of serum is limited, as in pediatric cases, the procedure is modified as follows to accommodate a 25 1.11sample. To a 1 ml centrifuge tube add 200 ~1 of 0.033 M sulfuric acid and 25 1.11of 5% sodium tungstate. While mixing on a vortex mixer add 25 ~1 of serum and continue mixing for about 3 minutes. Centrifuge and remove 2 ,ul of the clear supematant for direct injection into the instrument. A linear response of peak area to serum uric acid concentration was obtained up to at least 25 mg/dl for both procedures. The instrument sensitivity is
311
such that as little as 100 pg of uric acid can be determined quantitatively. In principle, it is therefore possible to reduce the volume of serum required for the protein precipitation step in the above micro procedure to about 0.25 ~1. The protein precipitation is not difficult to carry out on this scale; however, it is also practical to dilute the serum lo-fold with distilled water and inject it directly onto the column. We have devised a short precolumn to prevent clogging the narrow analytical column with protein. The precolumn can be easily replaced every few days, thus making deproteinization unnecessary. This approach works well and for routine situations the improvement in convenience is a decided advantage over previous methods. At present we are able to analyze 20 samples per hour on the chromatograph. The instrument is interfaced to a Digital Equipment Corporation S/E computer which automatically acquires and processes the data. The program was modified according to the recommendations of Chilcote and Mrochek [ 121 and is written in SABRE and FORTRAN 2. The details of the interface and a listing of the program are available on request.
AutoAnalyzer
procedures
Standard single channel AutoAnalyzer methods were used to corroborate the present results. The procedures used at both sites were standard colorimetric methods based on the reduction of phosphotungstic acid to tungsten blue. At one location (Upjohn Company) the classical chemistry of Brown [ 131 was used incorporating sodium cyanide in the dialyzed serum presumably to intensify the color and minimize the possibility of turbidity. This variation was not used at the other location (Erie County Laboratory) [14]. Results
and Discussion
The liquid chromatography of serum uric acid is essentially the same as that reported earlier for urine samples [lo]. The need for selectivity is more severe in the urine assay due to the higher concentrations of oxidizable metabolites. Combining the resolution of high performance liquid chromatography with the selectivity of electrochemistry results in great overall specificity for uric acid. The interferences normally implicated in the phosphotungstic acid method (i.e. ascorbic acid, xanthines, cysteine, glutathione, L-DOPA, and gentisic acid) do not contribute significantly to the present method even when present at lOO-fold the normal physiological concentrations. Corroboration of the specificity for uric acid was also carried out by enzymatic (uricase) removal of uric acid from a serum pool [15] prior to analysis by LCEC. No detectable signal was obtained for samples pretreated in this way. The precision obtained for repeated analysis of 0.5 ml and 25 ~1 aliquots of a 6.2 mg/dl serum pool was 1.7% and 1.8%, respectively. A correlation coefficient greater than 0.99 was obtained for both collaborative AutoAnalyzer studies using the phosphotungstic acid method. Fig. 1 illustrates the data for serum samples assayed by both LCEC and the calorimetric method which includes cyanide. Liquid chromatography has not yet been recognized as a routine tool for clinical analysis and it would therefore be difficult for many laboratories to
Fig.
1.
Correlation
between
liquid
chromatography
(y)
and
AutoAnalyzer
(I’)
procedures
for
104
serum
samples.
adopt the present method. Nevertheless, this situation is likely to change in the near future. The manifest advantages of LCEC may well prove it to be an attractive alternative to calorimetry in some cases. Acknowledgement The authors are indebted to Dr Terry J. Gilbertson (Upjohn Company, Kalamazoo, Mich. U.S.A.) and Dr David C. Wenke (Erie County Laboratory, Buffalo, N.Y.) for collaboration on the AutoAnalyzer methods. This investigation was supported by a grant from the National Science Foundation (GP42452X). References 1
R.
2
M.G.
Wells,
Clin.
Martinek.
J. Am.
Med.
Chim.
3
L.G.
Mm-in,
Clin.
4
N. Gochman
and
5
S. M&es.
C. Thompson
6
R.J.
and
Troy
7
G. Park,
8
P.A.
9
W.N.
R.N.
Kelly
Chem.. J.M.
W.C. Clin.
and
Kissinger.
and
Purdy, and
(1974) R.W.
W.R. 16
L.J.
Felice,
11
G. Dryhurst,
12
D.D.
13
H.J.
Brown,
J. Biol.
14
R.J.
Henry,
C. Sobel
15
D.G.
Remp,
Forts&r. and
Stand.
J.E.
Chim.
and
Forsch..
J. Kim,
Methods
Acta, Anal. Chem.,
Riggin.
158 Clin.
17
Clin.
(1971)
1154
Chem.,
27
20
(1970)
L&t..
(1974)
790
401
5 (1972)
887
191
Clin.
Mrochek,
Chem.,
Chem.,
White,
Chem.
233
379
Roach,
(1970)
R.M.
(1970)
51
Clin.
J. Wyngaarden.
P.T.
32 (1968)
Clin.
Chem.,
10
Chilcote
20
22
Schmitz.
Adams
Simkin,
Technol.,
Acta.
16
L.A. 34
(1972)
Clin.
Chem.,
(1945)
601
Am.
J. Clin.
Chem..
(1970)
Pachla 17
D.C.
Wenke,
Clin.
Chem.,
47 (1971)
Path&
6 (1970)
707
and
751 28
1
(1957)
152
and
645
20
(1974)
992
