CLIN. CHEM. 41/10, 1509-1517
(1995)
#{149} General
Clinical Chemistry
and Plasma Glutathione Measured in Healthy Subjects by HPLC: Relation to Sex, Aging, Biological Variables, and Life Habits Blood
France
Michelet,’
Ren#{233} Gueguen,’
Pierre Leroy,2 Maria Wellman,3
We report an HPLC method for measuring the concentrations of reduced (GSH) and total (GSHt) free glutathione in human plasma and whole blood. The chromatographic step was coupled with a postcolumn derivatization reaction and fluorometric detection. The linear range was 0.81-13.02 mol/L, and the detection limit was 0.13 mol/L. In healthy adults (ages 18-73 years), mean concentrations were 941 ± 155 mol/L for GSHt and 849 ± 63 mol/L for GSH in blood (107 men, 94 women), and 3.39 ± 1.04 mol/L for GSH in plasma (66 men, 58 women). Blood GSHt but not GSH was significantly lower in children (32 boys, 32 girls: 872 ± 157 p.mol/L) than in adults. Blood GSHt and GSH appeared to be correlated positively with the number of cigarettes smoked per day and the regular practice of physical exercise, and negatively with alcohol abstinence. We observed positive correlations between blood GSHt and cholesterol and calcium concentrations, and between blood GSH and cholesterol concentration. Indexing Terms: biological variation/y-glutamyltransferase/antioxidants/cholesterol/free radicals/sex- and age-related effects The tripeptide glutathione (gamma-L-glutamyl-Lcysteinylglycine) is the major intracellular nonprotein thiol compound and plays a major role in the protection of cells and tissue structures (1). This tripeptide acts as a nucleophilic scavenger and as an enzyme-catalyzed antioxidant in electrophilicloxidative tissue injuries that arise from both endo- and exogenous sources. Intracellular and blood glutathione concentrations are in the millimolar range, whereas plasma concentrations are in the micromolar range. The tripeptide can be free (GSHt), reduced (GSH), oxidized (GSSG), or bound to proteins.5 Glutathione allows the detoxification of free radicals and oxygen-reactive species involved in such diseases as atherosclerosis, rheumatoid arthritis, adult respiratory distress syndrome, or reoxygenation injury; consequently, its concentration decreases in such diseases (2). A decrease in erythrocyte GSH content has been 1
Centre de M#{233}decine Preventive,
54500 Vandoeuvre 2
Laboratoire
les Nancy,
de Chimie
2 avenue
and Gerard
Siest1’3’4
reported in patients affected by y-glutamylcysteine synthetase deficiency and hemolytic anemia (3). Decreased hepatic GSH concentrations have been observed in alcoholic (4, 5) and nonalcoholic (4, 6) liver diseases. Plasma GSHt concentrations also decrease in such diseases (7). Clinical studies and laboratory investigations of patients infected by HIV suggest that this syndrome may be the consequence of a virus-induced cysteine deficiency that results in a decrease in glutathione in plasma (8, 9), lung epithelial lining fluid (8), mononuclear cells (9), and T lymphocytes (10). High glutathione concentrations, on the other hand, may be a disadvantage for cancer treatments, as glutathione is involved in some of the chemotherapy resistance processes (11,12). In healthy subjects (13-28), mixed results have been reported; GSH concentrations vary from 684 (16) to 2525 moVL (13) in blood, and from 2.22 (22) to 11.36 j.mol/L (25) in plasma. Differences were often correlated with gender and (or) age, but the small size of the studied populations, especially when divided into subgroups, did not allow reliable statistical analysis. Moreover, no analytical reference method has been defined, either for sample preparation or for glutathione measurement. Therefore, the reference interval and the physiological variations of its concentration have to be determined. Consequently, our aim was, first, to develop a method for reproducible measurement of GSHt that would allow an accurate study of its redox status in biological samples. Then, we determined GSH and GSHt concentrations in blood (265 subjects), as well as plasma GSH (148 subjects). Results were analyzed in relation to gender and age and to some life habits such as tobacco or alcohol consumption, physical exercise, and oral contraception. Finally, we studied the correlations between glutathione concentrations and other biological variables usually determined to estimate individuals’ health, i.e., glucose, calcium, cholesterol, and triglyceride concentrations, and y-glutamyltransferase (GGT) activity.
du Doyen Parisot,
France.
Analytique,
Alain Nicolas,2
Facult#{233}de Pharmacie,
54000 Nancy, France. Centre du M#{233}dicament, URA CNRS 597, 30 rue Lionnois, 54000 Nancy, France. Author for correspondence. Fax 33/83 32 13 22; e-mail
[email protected]. I Nonstandard abbreviations: GSH, reduced glutathione; GSSG, oxidized glutathione; GSHt, total free glutathione; GGT, y-glutamyltransferase; OPA, o-phthalaldehyde; DTT, 1,4-D,L-dithiothreitol; and LOD, limit of detection. Received January 24, 1995; accepted June 19, 1995.
Materials Subjects
and Methods
The human population studied included 265 apparently healthy subjects, 139 males (107 adults, median age 40.0 ± 11.9, and 32 children, median age 13.5 ± 1.9) and 126 females (94 adults, median age 38.3 ± 13.2, and 32 children, median age 13.3 ± 1.9), voluntarily visiting the Center for Preventive Medicine of Vandoeuvre-les-Nancy (France) for a general health CLINICAL
CHEMISTRY,
Vol. 41, No. 10, 1995
1509
examination. None of these individuals was using any drug, except that some women (n = 27) were taking oral contraceptives. Data regarding tobacco, alcohol consumption, and drug intake were obtained by ques-
equipped with a 150-W xenon short-arc lamp. All data collection and calculations were by integrator (Chromjet; Spectra-Physics).
tionnaires
Analytical Methods
from
individuals
over
18. The
procedures
followed were in accordance with the ethical standards of the Centre de M#{233}decine Preventive’s responsible committee. Materials Reagents.
GSH,
GSSG,
o-phthalaldehyde (OPA), (DTT) were purchased from Fluka (Buchs, Switzerland); n-decyl sulfate sodium salt was from Eastman Kodak (Rochester, NY); perchloric and boric acids, disodium EDTA, and Brij-35 were from Merck (Darmstadt, Germany); potassium dihydrogen phosphate, NaOH, and HC1 were from Prolabo (Paris, France); ethanol was from Farmitalia Carlo Erba (Milano, Italy); serine was from SigmaAldrich (Saint-Quentin Fallavier, France); and acetonitrile was from SDS (Peypin, France). All chemicals and solvents were of analytical-reagent grade and were used without further purification. Solutions. The perchloric acid solution was obtained by dilution of concentrated perchloric acid (700 g/L) in an equal volume of HPLC-grade water and addition of 2 mmollL disodium EDTA. The L-serrne borate solution was prepared in phosphate-buffered saline at a final concentration of 100 mmol/L. The OPA solution was prepared daily as follows: 5.3 g of boric acid and 7.5 mL of 300 g/L NaOH solution were decanted into a 250-mL volumetric flask and dissolved in -200 mL of HPLCgrade water. OPA (0.2 g) was dissolved in 10 mL of 950 mL/L ethanol and added to the previous solution, together with 1 mL of a 300 mIJL Brij-35 solution. The final volume was adjusted to 250 mL with HPLC-grade water. The solution was degassed by ultrasonication for 5 mm before use and discarded if not used within 24 h. No increase of background signal was observed within a day. Stock solution of GSH was prepared daily at a concentration of 0.1 g/L (325 moWL) in 0.1 mol/L HC1 containing 2 mmol/L disodium EDTA, and kept in the dark at 5 #{176}C. Apparatus. The HPLC system consisted of a double reciprocal piston pump fitted with a membrane pulse damper (Spectroflow 400; Applied Biosystems, Foster City, CA), an injection valve equipped with a 50-p.L sample loop (Rheodyne 7125; Rheodyne, Cotati, CA), and a column oven (Croco-Cil; Spectra-Physics, Les Ulis, France). The HPLC had a guard column (4 X 4 mm) and an analytical column (125 X 4mm) prepacked with LiChrospher 100 RP-18 and end-capped (5 .tm; Merck-Clevenot, NogentlMarne, France). The column outlet was connected to a three-way postcolumn reagent delivery pump (Hitachi 655 A-13; Merck-Clevenot). Derivatization took place in a knitted Teflon open-tubular reactor (3 m X 0.5 mm; Model 5-9206; Supelco, Saint-Quentm Fallavier, France) at room temperature. The fluorescence detector (Shimadzu RF-551; Touzart et Matignon, Vitry-sur-Seine, France) was
and 1,4-DL-dithiothreitol
1510
CLINICAL CHEMISTRY, Vol. 41, No. 10, 1995
Blood sample collection and treatment. Venous blood samples (5-mL) were collected in EDTA-Vacutainer Tubes (Becton Dickinson, Grenoble, France) between 1145 and 1230, and blood was kept at room temperature before treatment. L-serine borate, 100 mmol/L (100 ML), was added within 15 ± 5 mm of blood collection. Then, 1 mL of blood was added to 0.5 mL of the perchloric acid solution and centrifuged at 10 000g for 3 mm at room temperature. The resulting supernatant was frozen in liquid nitrogen and stored at -80 #{176}C before analysis (within 15 days). Plasma was obtained from the residual blood by centrifugation at 60g for 10 mm at 4 #{176}C. Then, within 25 ± 5 mm of blood collection, 1 mL of plasma was mixed with 0.5 mL of the perchloric acid solution. The same procedure as above was performed. Before analysis, samples were thawed in a water bath at 37 #{176}C, and a 50-L aliquot was injected onto the column for GSH assay. Acidic supernatants from blood samples were usually diluted 100fold in a solution of 0.1 mo]IL HC1:2 mmol/L disodium EDTA, whereas acidic supernatants from plasma samples were injected without further dilution for GSH measurement. Concerning blood GSHt, a 100-p.L aliquot of the acidic supernatant was reduced by addition of 100 L of NaOH (4 mol/L) and 100 L of DTT (50 mmolfb) prepared in 0.1 mol/L HC1. After 20 mm at room temperature in the dark, 100 j.L of the resulting mixture was diluted 50-fold in the HC1 solution, and a 50-L aliquot was injected onto the HPLC system. Determination of GSH and GSHt. HPLC measurement of GSH was realized with the method of Leroy et a!. (29), which is based on the reaction between the sulfhydryl and amino groups of GSH and OPA. Briefly, the mobile phase was 90% (by vol) 0.5 mmoL1L n-decyl sulfate sodium salt, 0.1 nimol/L disodium EDTA, and 10 mmol/L phosphate buffer, pH 2.5, and 10% (by vol) acetonitrile. Flow rates of mobile phase and postcolumn derivatizing reagent were 1.2 and 0.3 mL/min, respectively. Separations were at 40 #{176}C, and the retention time of GSH was 4 measurements of the same sample) had to be 98%; thus we decided to quantify GSH and GSHt with the same calibration curve. GSH and GSHt concentrations in a given sample were calculated on the basis of peak areas according to the calibration curve. Blood sample stability. The stability of the acidic supernatant was tested on 479 days (Fig. 1). Measurement on day 0 was done immediately after acid precipitation and centrifugation, without the freezing step (see Materials and Methods). Aliquots were then frozen, and three of them were thawed for each of the nine subsequent measurements. We determined concentrations of 1187 ± 72 mol/L for GSH and 1263 ± 86 tmol/L for GSHt. Thus, CVs of the 10 measurements were 6.0% for GSH and 6.8% for GSHt. These CVs reflect the cumulative effects of sample preparation, analytical factors, and storage. No trend being observed (Fig. 1), we concluded that there was good mol/L)
stability of the acidic supernatants after blood samples were prepared conditions.
Physiological
Variations
stored at -80 #{176}C under the proposed
of Glutathione
Distribution of glutathione concentrations. The hypothesis of normal distributions for blood (n = 265) GSHt (Fig. 2A) and GSH (Fig. 2B) and plasma GSH (n = 148, Fig. 2C) was acceptable. Respective P values were 0.060, 0.328, and 0.137. Sex and age influence. No significant difference was apparent between glutathione concentrations determined in the male population and those in the female one, neither in blood (Table 1) nor in plasma (Table 2).
2000
1500 0
E a, C
0
1000
500
o GSH #{149} GSHt
00
100
200
300
400
500
Days of storage Fig. 1. Glutathione stability in acidic supematants. Sample was prepared from blood as described in Materialsand Methods and aliquots were stored at -80 c except for the first determination, which was realized without the freezing step. Glutathione concentration was then determined after 4-479 days of storage. Values are means ± SD of three determinations for each of the 10 measurements. CLINICAL
CHEMISTRY,
Vol. 41, No. 10, 1995
1511
500
600
700
800
900
100011001300
1300
1.25 1.75 2.252.753253754.254755.255.756256757.25
1400
7.75 8.25
Plasma GSH, ijmoIII.
Blood GSHt, pmol/L 80
B 70
60 50
40
30 20 10
Fig. 2.
Normal distribution of glutathione concentration: (A) blood GSHt; (B) blood GSH; (C) plasma GSH. The hypothesisof normal glutathione distribution was tested by the KH12 test; respective P values obtained were 0.06, 0.33, and 0.14.
0 400
500
700
600
800
900 1000 1100 1200 1300 1400
Blood GSH, pmol/L.
The mean blood glutathione concentrations were compared in four age groups without sex distinction: 10-17, 18-39, 40-59, and 60-73 years. Blood GSHt values varied significantly with age, as shown by the statistical difference between the youngest subjects and the adult population (Table 1). Concentrations of blood GSHt in adults and blood GSH in the whole population did not differ from one group to another. Consequently, we determined blood GSH and GSHt concentrations of 849 ± 163 moI/L and 941 ± 155 .tmoI/L in adults; the respective concentrations in children were 814 ± 171 imo1/L and 872 ± 157 mo1IL. Analysis of GSHJGSHt ratio during aging showed a slight decrease in the ratio; however, the difference was significant only between the oldest and youngest groups (Table 1). The same analysis was realized with plasma GSH (Table 2); concentrations declined slightly Table 1. Blood GSHt and GSH concentrations
(mol/L)
Male
and their ratio (%) according to sex and age.
Female
Age, years
n
GSH
GSH/GSHt
n
10-17
32
862
± 166
808 ± 183
93.5 ± 8.9
32
± 14.4a
GSHt
with age, and the decrease was significant between the extreme groups. No significant difference appeared in the adult populatmon. Tobacco influence. We defined three groups of subjects according to their smoking habits: nonsmoker, ex-smoker, and smoker. Blood GSHt and GSH concentrations differed significantly between nonsmokers and heavy smokers (Table 3). Blood GSHt and GSH concentrations were positively correlated with the number of cigarettes smoked per day within the smoker group (Table 4). The same pattern could be observed in plasma; the highest GSH concentration was measured in the subjects smoking >20 cigarettes per day, but differences did not reach statistical significance (Table 3). Alcohol consumption effect. Blood GSHt and GSH concentrations were lower among abstemious subjects
GSHt 882
Whole populatIon
OSH
GSH/GSHt
820
±
158
92.9
±
± 158
832
± 167
90.4
± 9.9
18-39
52
964
881 ± 150
91.3 ± 7.7
56 921
40-59
45
937 ± 162
834 ± 178
88.9 ± 10.3C
28
910 ± 138
846 ± 133
60-73
10
1017 ± 1400
882 ± 127
87.4 ± 10.9
10
968 ± 171
830
18-73
6.7
148
±
± 206
93.3 ± 85.8 ±
n
GSHt
GSH
64
872 ± 157
814 ± 171
93.2 ± 7.9
942
856
90.8 ± 8.9
108
±
153
± 161
838 ± 162
9.7
73
927 ± 154’
141d
20
993 ± 158#{176} 856 ± 173
GSWGSHt
90.6 ± 10.3 86.6 ± 12.6”
107 958 ± 153 861 ± 162 89.9 ± 9.3 94 923 ± 155 836 ± 163 90.7 ± 10.6 201 941 ± 155 849 ± 163 90.3 ± 9.9 Means of these age groups are significantly different from the mean of the youngest group, with P-values of: a 0.0044, #{176} 0.0123, C 0.045, d 0.0384, e 0.049,
‘0.0398,
0.0040, and
“
0.0071.
1512 CLINICAL CHEMISTRY, Vol. 41, No. 10,
1995
concentration
than among alcohol consumers (Tables 3 and 4). No difference appeared among people whose daily alcohol consumption was >50 g, but the number of individuals in this group was quite small (n = 22). Plasma GSH
was not affected by alcohol consumption (Tables 3 and 4). Physical exercise. Practice of physical exercise was associated with increased blood GSHt and GSH concentrations (Table 4), but with the difference being significant only for GSH (Table 3). Moreover, there was a significant decrease of glutathione oxidation, as shown by the increase of the GSH/GSHt ratio, when subjects practiced regular physical exercise (Table 3). Results did not reach statistical significance in plasma (Tables 3 and 4). Oral contraceptive use. There was a tendency towards decreased values among women who used oral contraceptives, but blood GSHt and GSH means were not significantly lower in this group (Tables 3 and 4). GGT activity. Our results did not show any correlation between GGT activity and glutathione concentra-
Table 2. Plasma GSII concentration (gLmol/L) according to sex and age. Male
Female
Wh ole population
Age,
years
n
10-17
13
OSH
n
3.39 ± 0.73
11
3.79
18-39 34 3.57 ± 0.87
36
3.35
± 1.40 ± 0.35a ± 1.09
17
3.24
5
3.00
40-59
25
3.49
60-73 18-73
7 66
2.96 3.47
GSH
n
± ± ± ± ±
0.69
24
3.57
GSH
±
0.98
70
3.46
± 0.93
0.99 0.91
42
3.39 ± 1.26
12
2.98 ± 0.65#{176}
074b
58 3.29 0.98 124 3.39 ± 1.04 Means are significantly different from each other at P = 0.0076a and
0.0276.”
Table 3. Influence of some life habits on blood GSHt and GSII concentrations
(JLmol/L)inadults.
Blood Variable
n
Nonsmokers and ex-smokers Nonsmokers Ex-smokers Smokers 1-9 cigarette(s)/day 10-19 cigarettes/day
138
20
Regular
physical
physical
activity
No oral contraceptive
use use
between
GSH/GSHt,%
n
OSH
90.1 ± 8.9
85
3.47 ± 1.08
90.6
± 8.6
51
3.45
823 ± 147
89.0 ± 9.3
34
3.50 ± 1.12
±
160 163C 137
865 813 871
±
90.4 91.0
939
±
822 879
±
± ±
912 ±
87 22
970 ± 151d 951 ± 148 940 ±
155d
153 148
± 171
±
r
339
#{247} -
± 1.14 3.51 ± 0.82 Not determined
± ±
±
11.1#{176}
77
92.3
±
36
90.2 ±
5.7#{176} 10.8
0.0192,
±
125
91.6
6.9
‘
0.01 22,
0.0184, ‘0.0327, and
±
GSHt, blood and plasma GSH concentrations,
-0.014 -0.017
Calcium GGT activity
No physical activity Physical activity No alcohol 1-20g/day alcohol 20g/day alcohol No smokers Smokers
Interval
r
± ±
113
3.34
#{176} 0.0111.
and other variables In adults. Plasma OSH
95% confidence
r
interval
95% confIdence
interval
+0.030
-0.148,
-0.189, +0.088 -0.186, +0.091
-0.063 -0.140
-0.237, +0.115 -0.309, +0.038
+0.111
-0.031,
+0.247
+0.075
-0.103,
+0.249
+0.071
-0.069,
+0.207
-0.036
-0.212,
+0.142
+0.075
-0.156#{176}
-0.295,
-0.010
-0.066
-0.248,
+0.122
+0.216 -0.038 +0.290 +0.169 +0.115 +0.211 +0.088
+0.156” -0.157#{176} +0.158#{176} -0.002 +0.004 +0.075 -0.087
+0.010, -0.289, +0.020, -0.140, -0.135, -0.065, -0.223,
+0.295 -0.018 +0.290 +0.137 +0.143 +0.212 +0.053
+0.066 -0.012 +0.010 +0.004 +0.042 -0.104 +0.060
-0.122, +0.248 -0.188, +0.166 -0.168, +0.186
+0.168#{176}
+0.029,
+0.300
+0.000
-0.162
-0.352, -0.043
+0.161#{176}
+0.022,
-0.153, +0.125 -0.156, +0.122
-0.052 -0.049
+0.153”
+0.012,
+0.287
+0.067
-0.072,
+0.204
-0.072
-0.295.
+0.072#{176} +0.010, -0.289, +0.158” +0.020, +0.031 -0.140, -0.024 -0.135, +0.074 -0.065, -0.052 -0.223,
±
No determined
Blood GSH
95% confidence +0.045,
7.4 10.8 9.1
89.1
794
+0.183”
No. of cigarettes/day Use of oral contraceptives
64
±
142’
843 ± 167
Cholesterol Triglycerides Glucose
Ex-smokers
90.6 90.3 90.7
165’
871 ± 152
blood
±
±
935 ± 150
C
±
3.11 3.11 3.72 3.38
±
51
0.0229, “0.0275,
3.21
11 21 6 60
837 890
27
a
38
12.3
89.0 ± 8.8
964
±
11.4
±
89.8 ± 12.2
162#{176} 163 158
0.93 0.50 0.99 1.15 0.94
±
846 ± 161
Blood GSHt
Variable
179 184
±
61
=
± 1.06
849 ± 155
150
92
Values are significantly different from each other at P
Table 4. Correlations8
OSH 840 ± 1 53b
±
1036 ± 157
121
activity
Oral contraceptive
927 957 893 969
13
1-5oglday alcohol 50 g/day alcohol No regular
938 ± 154
47 62 28
consumption
± 153
91
21
cigarettes/day
No alcohol
GSHt
934
Plasma
+0.314
+O.l73
+0.029,
-0.214”
-0.398, -0.011
+0.305
+0.292
-0.173,
+0.205
+0.181
-0.137, +0.218 -0.276, +0.076 -0.120, +0.235 -0.178,
+0.178
Not determined
Pearson correlation coefficient (30). “Significant correlation (P
