Genetic Dysregulation of Glutathione Synthesis Predicts Alteration of Plasma Thiol Redox Status in Schizophrenia

Subjects

All 54 patients (average age, 36.7 ± 11.1; 39 men, 15 women; duration of illness, ≥15 years) and 64 controls (average age, 35.7 ± 11.8; 35 men, 29 women) were recruited with fully informed written consent according to ethical guidelines of the Lausanne University. They were all of Caucasian origin and were assessed by using the Diagnostic Interview for Genetic Studies (DIGS), developed by the NIMH (42). Subjects with a major mood, psychotic, or substance-use disorder or a first-order relative with a psychotic disorder were excluded as controls. Patients, recruited from the in/outpatient schizophrenia units of the Psychiatry Department, Lausanne University Hospital, met the criteria for schizophrenia or schizoaffective disorder of the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV). Patients with a history of neurologic illness or head trauma, with mental retardation (IQ less than 60), or with a diagnosis of drug/alcohol dependence were excluded. Eight patients were untreated, and 46 were receiving antipsychotic medication (40 atypical, six typical) at therapeutic doses. To evaluate the effect of medication, it was computed to risperidone equivalents, as the majority of patients were treated with atypical antipsychotics (65).

Blood preparation

Blood was collected by venipuncture between 7 and 8:30 AM under restricted-activity conditions and fasting from the previous midnight. Then 20 ml of blood was allowed to drop in Vacutainer-tubes coated with Li-heparinate (Becton Dickinson), which were previously placed on ice. The blood was immediately centrifuged at 3,000 g, 5 min, 4°C; the pellet, corresponding to blood cells, was washed 2 times with 0.9% NaCl and frozen at −80°C until analysis. The supernatant, corresponding to the plasma, was recovered, sampled in aliquots, and kept at −80°C until analysis.

Amino acid analysis in plasma

The following free amino acids were quantified in plasma, as described in (51): taurine, hydroxyproline, threonine, serine, aspartate, glutamate, glutamine, proline, glycine, alanine, citrulline, valine, cystine, methionine, isoleucine, leucine, tyrosine, phenylalanine, ornithine, lysine, histidine, arginine. In brief, deproteinization was achieved by adding 50 μl 5-sulfosalicylic acid (160 g/L) containing the internal standards to 200 μl plasma. After mixing and standing (10 min), the sample was centrifuged 4 times at 14,600 g, 15 min, and the supernatant was kept at −80°C until analysis. Then 100 μl was injected into an amino acid analyzer (Biochrom, 30+ Model), and amino acids were detected by postcolumn reaction with nihydrine. Analytic performance was assessed for all amino acids by internal (IQC, home made) and external quality controls (ERNDIM; http://www.erndimqa.nl/). Typical coefficient of variation [CV (%), N = 75] were <3% for all amino acids except for Cit (<4%) and because of low concentrations in IQC sample for Asp (<18%; mean, 3 μM) and for Hyp (<22%; mean, 13 μM).

Thiols analysis in plasma

The following thiol containing amino acids and peptide were quantified in plasma as described before (62): total (protein-bound forms, free reduced and free oxidized form) cysteine, homocysteine, and cysteinyl-glycine. In brief, the thiols were reduced and/or decoupled from proteins by reaction with tris(2-carboxyethyl)phosphine. After deproteinization with perchloric acid, the thiols were derivatized with 7-fluorobenzofurazane-4-sulfonic acid (SBD-F). The SBD-F derivatives were analyzed by HPLC followed by fluorometric detection.

Analytic performance was assessed for total homocysteine by internal (IQC, home made) and external quality controls (ERNDIM; http://www.erndimqa.nl/). Typical coefficient of variation [CV (%)] were <5.2% (at 5.5 μM; N = 120) and <3.3% (at 16.4 μM; N = 143). Total cysteine and total cysteine-glycine have higher CV (%), usually <10%. The levels of the oxidized form of cysteine, the free cystine levels in plasma were analyzed as described in the earlier section, “amino acid analysis.”

Cell cultures

Human fibroblast cultures were established from skin biopsies, as described (58). Cells were grown with DMEM medium [0.5 l medium with 10 ml Ultroser-G plasma, 5 ml penicillin/streptomycin, 5 ml sodium pyruvate (100 mM)] at 37°C in a humidified atmosphere containing 5% CO₂/95% air.

GSH content in fibroblast

Cultures of fibroblast (four petri dishes of 10 cm diameter, confluent) were collected after three passages. They were removed from flasks with trypsin, washed, resuspended in 4 ml phosphate buffer (0.1 M, pH 7.4) and sonicated. GSH contents were measured by a colorimetric GSH assay kit (Calbiochem, San Diego, CA). This method is based on a colorimetric assay of a chromophoric thione formed specifically between the reagent and GSH. The protein levels of fibroblasts were determined by using the Biorad Kit with the Advanced Protein Assay reagent.

Genotyping

DNA was purified from blood samples by using NucleonBACC3 system (Amersham Pharmacia Biotech). GCLC GAG TNR polymorphism was assessed as originally described (63) with the following modifications: PCR amplification was performed in reaction mixtures of 25 μl containing 40 ng of each primer, 200 μM of each NTP and 0.3 U EuroTaq DNA polymerase with its buffer (Euroclone, Pero, Italy). Temperature cycling of the PCR was as originally described. PCR products were separated on an 8% polyacrylamide gel (6 × 8 cm) for 5 h with 40 V and stained by using SilverXpress Silver Staining Kit (Invitrogen, Carlsbad, CA).

Statistical analysis

Statistical analysis was performed by using the SAS statistical package version 9 (SAS Institute, Inc., Cary, NC). Multivariate analysis of variance (MANOVA) was assessed by using the amino acid levels as dependent variables and sex or GCLC GAG TNR genotype as fixed factors, and age and medication as covariables (GLM procedure, SAS/STAT User's Guide, version 6, 1989). A similar analysis was made between controls and patients groups by using age as covariable. Differences between genotype distributions were tested by using χ² or Fisher two-tailed test (FREQ procedure). Correlations between protein levels were studied by using the CORR procedure. A type I error, p < 0.05, was used to reject the null hypothesis.

Thiols and amino acid levels in high- and low-risk GCLC GAG TNR patients and in control subjects

In the present sample, the GCLC GAG TNR polymorphism was differently distributed (p = 0.038) between 64 controls and 54 patients (genotypes 7/7, 7/9, 7/8, 8/9, 8/8 & 9/9: in controls, 33, 26, 2, 2, 0, and 1 vs. in patients 18, 19, 8, 4, 2, and 3). In patients, the polymorphism allowed us to separate 37 low-risk (genotypes 7/7 and 7/9) and 17 high-risk patients (genotypes 7/8, 8/8, 8/9, and 9/9). The two groups did not differ by age and sex. Low-risk patients had a mean age of 35.4 ± 10.1 years and a male-to-female ratio of 2.7, whereas high-risk patients had a mean age of 35.5 ± 11.6 years and a male-to-female ratio of 2.4.

Skin fibroblasts of high-risk patients had decreased GSH levels, as compared with fibroblasts from low-risk patients (Table 1). Considering age and sex as covariables, MANOVA analysis revealed that the thiols and amino acid pattern was different in high risk, as compared with low-risk patients (p = 0.029). Univariate analysis showed that high-risk patients had significantly lower levels of total cysteine (oxidized + reduced form + protein bound; minus 20.4%; p = 0.004), but higher levels of free cystine (+43.3%, p = 0.005), than low-risk patients (Table 1). Moreover threonine, glutamine, citrulline, and arginine levels were also significantly higher in high-risk genotype patients (Table 2).

The thiols levels were also compared between all schizophrenia patients and control subjects. Free cystine (+20.7%; p = 0.033) were higher in patients (+20.7%; p = 0.033) as compared with control subjects. Interestingly, as compared with control subjects (36.3 μM), plasma free cystine levels were similar in low-risk genotype patients (36.9 μM), whereas they were 46% higher in high-risk genotypes (52.9 μM). Likewise, plasma cysteine levels, compared with control subjects, were higher in high-risk but not different in low-risk genotypes (Tables 1 and 3). Furthermore, fibroblast GSH levels were lower in high-risk but not different in low-risk genotypes. In the control-subjects group, the cysteine and cystine levels did not differ between the 59 low-risk and five high-risk genotypes. For cysteine, they amounted, in micromoles per liter, to, respectively, 191.7 ± 46.2 and 188.0 ± 20.4, whereas for cystine, 35.6 ± 16.7 and 36.9 ± 13.5. Therefore, it is justified to pool them for comparison with the patients.

Influence of antipsychotic medication

Clinical studies have shown that antipsychotic medication can influence amino acid levels. To evaluate the effect of antipsychotics treatment on amino acid levels, the medication doses were computed into risperidone equivalents (65). Negative correlations were observed between citrulline levels (R = −0.315 p = 0.031 respectively) and leucine levels (R = −0.300; p = 0.041) with the medication doses. No dose effect of treatment was observed on all other amino acid and thiols levels. Moreover, no difference of antipsychotic dose was observed between high- and low-risk genotype patients (p = 0.11, ANOVA). Among the 40 patients who received atypical antipsychotics, 27 were with low-risk and 13 with high-risk genotypes. For the six patients receiving typical antipsychotics, five were with low risk and one with high-risk genotypes. No difference (Fisher test, p = 0.65) was observed in exposure of the patients to typical vs. atypical antipsychotic drugs in the high- vs. low-risk patient genotype groups. It is of note that there were eight untreated patients, five with low-risk genotype (13.5% of all low risk), and three with high-risk genotype (17.6% of all high risk). The distribution of these untreated patients between the two genotypes was not significantly different (p = 0.7; Fisher test). Interestingly, citrulline levels were increased in high-risk patients, as compared with low-risk patients. Comparing citrulline levels between high-risk patients (34.9 μM) and controls (29.1 μM) revealed a significant difference (p = 0.010) difference in exposure.

Plasma redox thiol status: methodologic considerations

Cysteine, its disulfide cystine, GSH and glutathione disulfide constitute the major low-molecular-weight thiol/disulfide redox control systems in mammals (30). However, in plasma, other aminothiols such as cysteine-glycine and homocysteine, also contribute to thiol redox status, although in much smaller proportion (37, 59). In plasma, cysteine is the most abundant aminothiol (total concentration about 200 μM); about 65% is protein bound, 30% is free oxidized (free cystine), and 3–4% is reduced. The plasma concentration of cysteine-glycine is less (total, ∼40 μM); 60% is protein bound, 30% is free oxidized, and 10% is reduced. At physiologic conditions, total homocysteine levels amount to ∼10 μM, which encompass the predominant protein-bound, the free oxidized fraction, and the small amounts of free reduced form (60). Most (∼65%) plasma glutathione (total concentration, ∼6 μM) is in the reduced form (3, 4, 37, 59), but its high intracellular concentration (40) makes the determination unreliable in the presence of hemolysis (1% hemolysis can result in an increase of ∼7 μM GSH in plasma). Therefore, GSH levels in plasma were not measured in the present study. Moreover, rapid oxidation or disulfide exchange reactions and redistribution of thiol species or both may also take place in freshly prepared plasma. In general, these factors were not taken into consideration (15): little information was given concerning free versus total thiols measurements and what precautions were taken to avoid artifacts. In the present study, although maximal care was taken to minimize these artificial factors during blood collection and processing (lowering temperature and pH), we thus favored, in a relatively difficult psychiatric clinical setting, the quantification of total thiols (cysteine, cysteine-glycine, and homocysteine) to their free reduced forms. Although precise redox states (E_h) of the cysteine/cystine, as imputed in experimental studies (31), could not be calculated, it still allows a robust and reliable overall picture of the redox thiol status in plasma.

GCLC genotypes related imbalance of plasma thiols

Schizophrenia is a complex dynamic disease with variations of symptoms over time (34). Plasma amino acid concentrations usually do not directly correlate with brain amino acid concentrations, and they are influenced by many parameters, including age, sex, medication, duration of disease, and lifestyle (21, 23, 57). In the present study, blood collection was done under fasting conditions and parameters like age, sex, and medication were taken into consideration. Moreover, all the patients were in stable and chronic situation with at least 15 years of illness. Alterations observed among patients only, according the GCLC GAG TNR high-risk versus low-risk genotypes, are probably less biased by confounding factors such as medication, illness duration, and lifestyle, and were analyzed by using age and sex as covariates.

High-risk GCLC GAG TNR patients represent around one third of all patients (17, 25). While the levels of total cysteine and of its free oxidized form, cystine, were similar between low-risk genotype patients and control subjects, high-risk genotype patients had 20% lower total cysteine and 43% higher free cystine levels as compared with low-risk genotypes. No alteration was observed with the other plasma thiols such as cysteine-glycine and homocysteine. Conditions causing oxidative stress have been reported to consume extracellular antioxidants, including vitamin C, protein thiols, bilirubin, urate, and α-tocopherol (26). However, reduced cysteine, the most abundant low-molecular-mass thiol component in plasma, represents an important extracellular antioxidant defense system. The decreased levels of total cysteine and increased levels of its oxidized form in high-risk genotype patients reflected enhanced oxidative stress processes or impaired redox regulation or both, as a consequence of the genetically impaired GSH synthesis. Combined with the decreased GSH levels in fibroblasts of high-risk genotype patients, these results show that compromised synthesis of the ubiquitous GSH system in patients with high-risk GCLC GAG TNR polymorphisms affected the thiols status in peripheral tissue, such as blood and fibroblasts. It should be noted that, in control subjects, both cysteine and cystine levels are not different between the two GAG TNR genotypes. Thus, in this respect, the patients carrying high-risk genotype are different from the high-risk controls. This is in line with the existence of additional risk factor(s) not yet identified in patients.

Other plasma amino acid changes

Plasma levels of other amino acids were also compared between high-risk and low-risk genotypes. High-risk patients had 16% higher glycine and 12% higher serine levels, as compared with low-risk patients. Glycine and serine have drawn particular attention in schizophrenia research. Some studies observed elevated plasma levels of glycine and serine in schizophrenia patients (5, 64), although others could not confirm these alterations (22, 47). These contradictory results may be explained by the difference of high-risk, low-risk genotypes distribution in these various schizophrenia patients' samples.

Compared with low-risk GCLC GAG TNR patients, high-risk patients had also increased levels of glutamine, citrulline, and arginine. Interestingly, increased glutamine levels have been observed in chronic schizophrenia patients (1), and increased levels of citrulline in drug-naïve patients (47). Moreover, as arginine and citrulline are, respectively, precursor and product in nitric oxide (NO) synthesis, their increased levels in high-risk patients might reflect higher NO metabolism. Higher NO levels or NO synthase activity were reported in red blood cells (27), platelets (14), and in postmortem brain (67) of schizophrenia patients. As excess of NO levels can lead to generation of the free radical peroxynitrite (6); this result further supports the involvement of nitrosative stress pathology in high-risk patients (35, 45). The negative correlation between citrulline levels and antipsychotic medication is consistent with the report that treatments with chlorpromazine, haloperidol, or clozapine normalize NO synthase activity (14).

Comorbidities

Apart from psychiatric symptoms, schizophrenia is known to have major somatic concerns like an increased risk for cardiovascular diseases (24), or for diabetes mellitus (29, 55), also in drug-naïve patients (48). Interestingly, all these three diseases were found to be associated with polymorphisms of GCLC and GCLM. The GCLC GAG TNR polymorphism was shown to be associated with schizophrenia and with diabetes (7, 25). The −129 C/T of GCLC and the −588 C/T of GCLM have been associated with an increased risk for cardiovascular events and myocardial infarction (32, 41).

In conclusion, based on clinical and preclinical studies, the GCLC GAG TNR polymorphism was proposed as a schizophrenia susceptibility factor. The present results show that the high-risk GCLC GAG TNR genotypes lead to alterations of plasma thiols levels that reflect a dysregulation of redox control, and an increased susceptibility to oxidative stress. This indicates that the high-risk GCLC genotypes are associated with a subgroup of schizophrenia characterized by redox imbalance. The plasma/fibroblast thiol levels could therefore potentially contribute to a biomarker profile useful for early diagnosis and monitoring the effectiveness of novel drugs targeting redox dysregulation in schizophrenia.