Introduction
The neurobiology of preclinical or ultra-high risk (UHR) psychosis states has been increasingly studied using neuroimaging and neurocognition techniques (Bartholomeusz et al., 2016; Chung and Cannon, 2015). Nonetheless, other lines of investigation, such as peripheral biomarkers, remain underexplored.Reference Zeni-Graiff, Rizzo and Mansur1–Reference Fusar-Poli, Cappucciati and Borgwardt4 Replicated evidence has indicated that oxidative stress imbalances are involved in the pathophysiology of psychosis.Reference Coughlin, Hayes and Tanaka5–Reference Magalhaes, Dean, Andreazza, Berk and Kapczinski7 However, mixed results regarding two of the primary enzymatic antioxidants, superoxide dismutase (SOD) and glutathione peroxidase (GPx), have been reported in schizophrenia populations.Reference Gonzalez-Liencres, Tas and Brown8–Reference Zhang, Chen and Tan10 Clinical studies focusing on the early stages of psychosis, wherein the outcomes are less affected by confounders, especially medications effects, have more consistently reported abnormalities in oxidative stress pathways.Reference Noto, Ota and Gadelha11–Reference Pedrini, Massuda and Fries13
The accumulation of reactive oxygen species (ROS), either through increased production or due to a relative lack of antioxidant defenses, leads to damage of biomolecules (ie, lipids, proteins, DNA),Reference Rizvi and Maurya14 which, in turn, has been implicated in disturbances in signal transduction and synaptic plasticity, among other processes.Reference Pedrini, Massuda and Fries13, Reference Dietrich-Muszalska and Kontek15 Overall, the regulation of oxidative stress has been proposed as an effector system of interest in UHR states.Reference Smesny, Milleit and Schaefer16 Notwithstanding, antioxidant enzymatic defenses have not yet been studied in populations that are UHR for psychosis.
The aim of this study was to compare SOD and GPx activity in a sample of subjects at UHR for psychosis with a healthy control group. Secondary aims were to verify the impact of possible associated factors (ie, mood symptoms, psychosocial functioning, body mass index, and history of childhood maltreatment) in the activity level of these enzymes.
Methods
This study was approved by the local ethics committee from Universidade Federal de São Paulo (Unifesp), and all the subjects above 18 years of age provided written informed consent prior their inclusion. For participants under 18 years old, an assent term was obtained and the written informed consent was provided by the legal responder.
Sample
Thirteen UHR subjects within ages 14–26 years were selected from those attending the PRISMA Early Intervention Program, a specialized outpatient unit for assistance and research in UHR for psychosis in São Paulo, Brazil.Reference Brietzke, Araripe Neto, Dias, Mansur and Bressan17, Reference Castro, Zanini, Gonçalves Bda, Coelho, Bressan, Bittencourt, Gadelha, Brietzke and Tufik18 All the subjects recruited for this study were seeking help for any kind of mental or emotional suffering and/or were referred to PRISMA by health professionals, including psychiatrists, psychotherapists, occupational therapists, counselors, and social workers. Approximately 1:10 patients referred to PRISMA fulfilled criteria to be considered UHR for psychosis, and all the potentially eligible subjects accepted to participate in this study. The composition and the recruitment method adopted at PRISMA were similar to those from other well-established cohorts.Reference Yung, Yuen and McGorry19 We used the criteria for UHR for psychosis proposed and validated by Yung et al, according to the classification of the Comprehensive Assessment of At-Risk Mental State (CAARMS) Scale.Reference Yung, Yuen and McGorry19 The criteria consist of three possible situations:
1. Attenuated positive symptoms: Presence of positive symptoms in moderate severity, but not fulfilling criteria for a psychotic episode; present more than 1 time per month for more than 1 hour per week, in the past year
2. Brief intermittent psychotic symptoms: Presence of brief episodes of a full psychotic illness, which might involve all of the symptoms of a psychosis (particularly delusions and hallucinations)
3. Trait and state risk factors: Vulnerable family history of psychosis in a first-degree relative or a diagnosis of schizotypal personality disorder associated with a decline in social and occupational functioning
For the control group, only healthy individuals were selected from the community, with the same age range and sex as the UHR group. Healthy volunteers had no current or lifetime history of any mental disorders according to the Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) Axis 1 (SCID-1). In addition, we included only subjects with no history of use of any psychotropic medication and negative family history of a major psychiatric disorder [defined as unipolar depression, bipolar disorder (BD), suicide, or psychosis] in a first-degree relative. The exclusion criteria for both groups were suicide risk, presence of acute or chronic general medical conditions associated with imbalances in oxidative pathways (ie, infections, cancer), or pregnancy postpartum period.
Clinical status was assessed by the following scales: Montgomery–Åsberg Depression Rating Scale (MADRS), Young Mania Rating Scale (YMRS), Global Assessment of Functioning scale (GAF), and Global Assessment of Symptoms Scale (GAS).
Blood collection
All blood samples were taken in the morning (between 8 am and 10 am) in the fasting state. After collection, blood samples were immediately placed on ice, allowed to clot for at least 30 min at room temperature, and centrifuged at 3200 rpm for 10 min. The obtained serum was aliquoted and then stored at –80°C for a maximum of 2 years, until the assay.
Superoxide dismutase and glutathione peroxidase assessment
The activity of SOD and GPx in serum was evaluated by the spectrophotometric method using enzyme-linked immunosorbent assay kits (Cloud-Clone Corp., Houston, TX, USA.). The samples were diluted (1:50 for SOD and 1:500 for GPx) and run in duplicate or triplicate. If the coefficient of variation between the replicate was higher than 10%, the samples were repeated. Measurements were recorded using a micro-plate reader at 450 nm.
Statistics
Statistical Package for Social Sciences (SPSS) Version 20.0 (IBM Corporation, Armonk, NY, USA) was used for the statistical analysis of data. The Shapiro–Wilk test was performed in all variables to assess normal distributions. The chi-square test was used for categorical variables. Correlations between continuous variables were made by Spearman’s rank correlation test. For unadjusted comparisons, we used the Mann–Whitney U test. For adjusted analysis, we used generalized linear models. Due to the positively skewed distribution of GPx and SOD activities, models with gamma distribution were used. Due to the nonlinearity of the models, the estimated β coefficients created by the model were transformed into rate ratio (RR) estimates. Differences were accepted as significant when α<0.05 (2-tailed).
Results
Clinical and demographic characteristics of the sample are described in Table 1. There were no statistically significant differences in sex, age, or ethnical background between the UHR and healthy control groups.
Table 1 Clinical and demographic characteristics of the sample
a Chi-square test;
b T-test;
c Mann-Whitney test.
UHR: ultra high risk; HC: healthy controls; SD: standard deviations; MADRS: Montgomery–Åsberg Depression Rating Scale; IQR: interquartile range; YMRS: Young Mania Rating Scale; GAF: General Assessment of Functioning; SOD: superoxide dismutase; GPx: glutathione peroxidase.
GPx and SOD activities were not correlated with age (r=0.236, p=0.236 and r=0.354, p=0.051, respectively); GPx was positively correlated with years of education (r=0.454, p=0.013), whereas SOD was not (r=0.261, p=0.208). GPx and SOD activity were not affected by sex (z=1.088, p=0.277 and z=0.080, p=0.936, respectively) or ethnicity (z=0.392, p=0.695 and z=0.029, p=0.977, respectively).
Unadjusted comparisons showed a lower level of activity of GPx in the UHR group, whereas there was no statistically significant difference in SOD. After adjusting for age and years of education, there was a significant lower activity of SOD in the UHR group compared to the healthy control group (RR=0.330, 95% CI 0.187; 0.584, p<0.001); the lower GPx activity in the UHR group remained significant (RR = 0.509, 95% CI 0.323; 0.803, p=0.004).
There were also positive correlations between GAF functioning score and GPx and SOD activities (r=0.399, p = 0.029 and r=0.427, p=0.030, respectively). There were no correlations between GPx and SOD activities and total MADRS score (r=–0.241, p=0.200 and r=–0.103, p=0.617, respectively) and total YMRS score (r=0.016, p=0.939 and r=–0.237, p=0.225, respectively).
Discussion
The results of this preliminary study indicate that, after adjusting for age and years of education, individuals at UHR for psychosis had lower activities of both SOD and GPx activity compared to the healthy control group. In addition, lower activity of GPx and SOD was associated with worse psychosocial functioning. This was the first study to assess GPx and SOD activities in an UHR for psychosis population, and provides potentially useful data for the study of oxidative imbalances in the preclinical stages of psychosis.
Indeed, although oxidative stress has a well-recognized role in the pathophysiology of schizophrenia and psychotic affective disorders, there is a scarcity of studies examining the effects of oxidative stress on at-risk populations. Recently, Coughlin et al Reference Coughlin, Hayes and Tanaka5 described a reduction in SOD activity in the cerebrospinal fluid. Only one study to date has assessed the cellular oxidative damage to nucleic acids in this population; the authors were not able to demonstrate differences between UHR and healthy controls.Reference Nordholm, Poulsen and Hjorthøj20
Studies assessing SOD and GPx in mental illnesses have provided heterogeneous results. The findings seem to be influenced by ethnicity, gender, stage of illness, medication, comorbidities, obesity, and smoking.Reference Flatow, Buckley and Miller21–Reference Mansur, Santos, Rizzo, Cunha, Asevedo, Noto, Pedrini, Zeni, Cordeiro, McIntyre and Brietzke25 Interestingly, this heterogeneity also appears in studies assessing the activity of these enzymes in individuals experiencing their first episode of psychosis.Reference Martínez-Cengotitabengoa, Mac-Dowell and Leza22, Reference Raffa, Atig, Mhalla, Kerkeni and Mechri23
Possible differences between chronic patients with schizophrenia and individuals transitioning to psychosis have not been directly evaluated. Nevertheless, indirect evidence could be found in one study that assessed travel-induced psychosis, in which the patients had higher SOD compared to controls.Reference Gong, Zhao and Yang26 In addition, studies assessing the effect of antipsychotics suggest that reduction in SOD commonly found in chronic patients could be due to medication.Reference Sarandol, Sarandol, Acikgoz, Eker, Akkaya and Dirican27, Reference Zhang, Zhou, Cao, Zhang, Wu and Shen28
This study has important limitations. We evaluated a relatively rare group of individuals, in one observation, with small sample size and therefore more susceptible to sampling error and type II error. The cross-sectional characteristics of this study do not allow the identification of causality effects.
Conclusion
The importance of early recognition and early intervention in the outcomes of individuals with schizophrenia is now well established.Reference Perkins, Gu, Boteva and Lieberman29, Reference Killackey and Yung30 It has been hypothesized that preventive interventions during putatively prodromal phases could delay, reduce severity, or primarily prevent the development of a psychotic illness.Reference Fusar-Poli, Bonoldi and Yung31–Reference McGorry, Nelson and Amminger33 There is, however, insufficient data to reliably identify who will and who will not transition to psychosis based solely on clinical characteristics.Reference Fusar-Poli, Carpenter, Woods and McGlashan34, Reference Tsuang, Van Os and Tandon35 Therefore, oxidative stress imbalances emerge as a promising candidate to be incorporated in future investigations of neurobiology of at-risk mental states.
