Introduction

Schizophrenia is a severe mental illness that typically begins in adolescence or early adult life. It is a heterogeneous disorder with diverse symptoms which are generally characterised as positive or negative [1,2]. Positive symptoms, such as delusions and hallucinations, are considered to be more transient; negative symptoms, such as blunting of affect and passive withdrawal, are regarded as more persistent and contribute more to impairment [3].

The dopamine hypothesis is one of the most influential theories about the etiology of schizophrenia. Initially, the hypothesis emphasised an etiologic role of hyperdopaminergia, but this was subsequently reconceptualised to specify subcortical hyperdopaminergia with prefrontal hypodopaminergia [4]. Normally, the prefrontal dopamine system suppresses the limbic dopamine system; however, in schizophrenia, this suppression seems to be reduced due to disturbed prefrontal dopaminergic activity, leading to elevated limbic dopaminergic activity [5]. The dopamine hypothesis is concordant with both the negative symptoms of schizophrenia, likely linked with prefrontal hypodopaminergia, and the positive symptoms, strongly related to enhanced limbic dopaminergic activity [6].

While psychopathology is traditionally explained by disturbed neurotransmitter function, polyunsaturated fatty acids (PUFAs) gained interest in terms of the etiology of structural and functional abnormalities of the developing nervous system in schizophrenia. PUFAs are major constituents of neuronal and myelin membranes but have also important functions in the regulation of neuronal migration, pruning, and synaptic plasticity. Evidence from animal studies shows that a profound experimental ω-3 PUFA deficiency can alter dopaminergic and serotonergic neurotransmitter systems. These associations between PUFA status and neurotransmission may explain the role of PUFAs on human brain function and behaviour [7]. PUFAs are also selectively concentrated in synaptic neuronal membranes and regulate vascular and immune functions that affect the central nervous system [8]. The ω-3 PUFAs have anti-inflammatory effects, suppress interleukin-1β, tumor necrosis factor-α and interleukin-6, whereas ω-6 PUFAs do not [9]. A positive correlation between ω-6 PUFAs and intracellular phospholipase-2 (inPLA₂) activity was observed in patients at ultra-high risk (UHR) for psychosis, while supplementation with ω-3 PUFA resulted in a significant decrease in inPLA₂ activity [10]. These potential pro-inflammatory effects of ω-6 PUFAs and anti-inflammatory effects of ω-3 PUFAs may be associated with psychopathology in different way.

The accumulated evidence suggests that patients with schizophrenia are deficient in key PUFAs [11]. However, it is not clearly understood which PUFAs are altered and how PUFA deficiencies relate to psychotic symptoms. Several studies have demonstrated an association between PUFA deficiencies and the severity of the negative symptoms of schizophrenia, whereas others have reported inconsistent findings [12–14]. It is also not known if PUFA alterations reflect risk factors such as nutrition, are a result of a biological predisposition to schizophrenia or if they are associated only with an existing psychotic state, where they might be influenced by disease related pathways such as oxidative damage to lipids [14]. As antipsychotic medication and also the progression of illness may alter PUFA levels, FA research focussed on drug-free first-episode patients to minimise this confound. Evaluations of PUFA profiles in individuals at the prodromal (UHR) stage of schizophrenia will help even more to clarify the role of FAs in schizophrenia etiopathogenesis.

Recently, α-linolenic acid (ALA), eicosapentaenoic acid (EPA), and ω-6 PUFAs were found to be significantly lower in young UHR individuals than in healthy controls [15]. This suggests that FA deficiency may be present prior to illness onset and may be useful as a potential risk biomarker for the high-risk state, using the classification of Davis [16]. To investigate state marker properties, the present exploratory study aimed to investigate the relationship between erythrocyte membrane FA levels and the severity of symptoms in individuals at UHR for psychosis. The association between membrane FAs and psychotic symptoms was also evaluated according to the predominance of positive or negative symptoms.

Methods

Results

Partial correlations between erythrocyte membrane PE fatty acids and symptom severity measures after controlling for age, sex, smoking, cannabis use, and scores on MADRS and GAF are presented in Table 1. The ω-6:ω-3 FA ratio was positively correlated with PANSS total and negative symptom scores and sum of the ω 3 fatty acids was negatively associated with PANSS negative symptom scores. Total PANSS scores were negatively correlated with two saturated FAs (myristic and margaric acids) and one ω-9 MUFA (nervonic acid) and positively correlated only with two ω-9 MUFAs (eicosenoic and erucic acids) and two ω-6 PUFAs (γ-linolenic and docosadienoic acids). PANSS positive symptom scores were significantly correlated with nervonic acid. In terms of PANSS negative symptom and general psychopathology scores, negative correlations were found with two saturated FAs (myristic and margaric acids) and one ω-9 MUFA (nervonic acid), whereas positive correlations were found with two ω-9 MUFAs (eicosenoic and erucic acids) and two ω-6 PUFAs (γ-linolenic and docosadienoic acids). One ω-3 PUFA (docosapentaenoic acid, (DPA)) was negatively correlated with PANSS negative symptoms scores. GAF scores were correlated with one ω-6 PUFA (docosadienoic acid).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Partial correlations coefficients between erythrocyte membrane phosphatidylethanolamine lipids and psychiatric measures controlling for age, sex, smoking, and cannabis use.

https://doi.org/10.1371/journal.pone.0151417.t001

Demographic and clinical characteristics of subgroups according to PANSS scores are shown in Table 2. No significant differences were found with regard to age; sex; use of alcohol, cannabis, tobacco, or any illicit drug; MADRS scores; or transition to psychosis. GAF scores were significantly lower in the negative-symptoms-dominant group than in the positive-symptoms-dominant group (p-value = 0.004, ANOVA post-hoc analysis).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Comparison of demographic and clinical characteristics of participants according to dominance of positive or negative symptoms.

https://doi.org/10.1371/journal.pone.0151417.t002

Table 3 include the results of group comparison of FA levels among the three PANSS subgroups. The sum of ω-3 PUFAs and myristic acid levels were significantly lower in the negative-symptoms-dominant group than in the positive-symptoms-dominant group. The sum of the ω-9 MUFAs and two single ω-9 MUFAs (eicosenoic and erucic acids), two ω-6 PUFAs (γ-linolenic and docosadienoic acids), and the ω-6:ω-3 FA ratio were all significantly higher in the negative-symptoms-dominant as compared to the positive-symptoms-dominant group.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Comparison of erythrocyte membrane phosphatidylethanolamine lipids levels according to dominance of positive or negative symptoms.

https://doi.org/10.1371/journal.pone.0151417.t003

Discussion

In this exploratory study of individuals at UHR for psychosis, the PE fraction of erythrocyte membrane FAs were associated with severity of psychopathology. Negative symptoms were negatively associated with saturated FAs (myristic and margaric acids), ω-3 PUFA (DPA), and nervonic acid but positively associated with several other ω-9 MUFAs (eicosenoic and erucic acids) and ω-6 PUFAs (γ-linolenic and docosadienoic acids). The sum of the ω-9 MUFAs and the ω-6:ω-3 FA ratio were both significantly higher, whereas the sum of the ω-3 PUFAs was significantly lower in subjects with predominant negative symptoms compared with those with predominant positive symptoms. No associations were observed between FAs and MADRS scores, suggesting that the associations between FAs and negative symptoms were not confounded by depression, a potential concern given their overlapping phenomenology. To the best of our knowledge, this is the first study conducted in neuroleptic-naïve UHR subjects that has demonstrated a relationship between membrane FA levels and the severity of symptomatology. The findings suggest that cell membrane lipid biology may be useful to assess during the onset phase of psychotic disorders and are consistent with the observation that supplementation with omega-3 PUFAs might be an effective and longer-term preventive treatment in this cohort of UHR patients [17,35].

Our findings of an association between membrane FA level and negative symptoms are comparable to earlier studies conducted in patients with schizophrenia. Glen and colleagues found that patients with predominantly negative symptoms exhibited lower levels of ω-3 PUFAs in red blood cells compared with those with persistently positive symptoms and controls [12]. Additionally, patients with low erythrocyte PUFA (sum of ω-3 and ω-6 FA levels) had more negative symptoms than those with high PUFA levels [13]. In a study in drug naïve-patients with schizophrenia, ω-3 PUFA deficiency was associated with negative symptoms [14]. Accordingly, it was demonstrated that supplementation with ω-3-rich fish oil might result in a significant improvement in negative but not in positive symptoms of patients with schizophrenia [36]. In a clinical trial of EPA supplementation in patients with schizophrenia, an increase in membrane ω-3 PUFAs was significantly correlated with change in PANSS negative symptom scores and total scores [37].

The prefrontal cortex plays an important role in higher brain functions. The basic activity of this brain region is considered to be the orchestration of thoughts and actions in accordance with internal goals [38]. It has also been implicated in planning complex cognitive behaviour, executive function, decision making, and moderating social behaviour [39]. It has been assumed that hypofunctioning of the cortical and prefrontal dopamine systems contributes to cognitive impairment and negative symptoms in patients with schizophrenia, such as diminished emotional expression and avolition [40]. The relationship between ω-3 PUFA deficiency and negative symptoms is concordant with the dopamine hypothesis of schizophrenia. In animal studies, deficiencies in ω-3 PUFAs have been found to cause reductions in levels of dopamine and D2 receptors in the frontal cortex [41,42] and a hypofunction of the prefrontal dopamine system, reflecting a correlate of negative symptoms and cognitive impairments [6]. In addition, ω-3 PUFA deficiency of rats affects dopaminergic neurotransmission and mRNA expression of D2 receptor genes in the frontal cortex [43–45].

Animal studies showed that oral intake of ω-3 FAs can increase and decrease the number of D2 receptors in the frontal lobe and striate body, respectively [46,47]. In a clinical trial of patients with first-episode schizophrenia, improvement of negative symptoms induced by ethyl-EPA supplementation was associated with metabolic brain changes, particularly increases in glutathione [48]. Glutathione is an antioxidant that protects dopaminergic neurons from oxidative and excitatory damage [49,50].

White matter changes including dysregulated myelination are presented in patients with schizophrenia [51]. Reduced levels of PUFAs in erythrocyte membranes may be also related to white matter integrity throughout various brain areas including the frontal lobe. An imaging study of patients with early psychosis found that more severe negative symptoms were associated with lower levels of integrity in the white matter associated with total PUFA or arachidonic acid (an ω-6 PUFA), whereas no significant associations were found with positive symptoms [52]. Furthermore, animal studies suggested that PUFAs are important for the development, structure, and functioning of neurons and glial cells [53] and for neurite outgrowth, exocytosis, and monoaminergic signal transmission [54]. Thus, PUFA deficiency may impair dopaminergic and glutamatergic neurotransmission in the frontal lobe, which is linked to the negative symptoms of schizophrenia [13]. These biochemical, imaging, pharmacological, and animal experimental studies support our findings.

In particular, nervonic acid, a major constituent of myelin membranes, helps to maintain white matter integrity [52,55]. In the present study, nervonic acid was negatively correlated with scores of the PANSS (total score and scores on the three subscales) in UHR individuals, which is consistent with earlier studies of patients with schizophrenia.

Myristic acid is an anchorage moiety, and plays a role in attaching proteins to lipid biomembranes. Linkage of myristic acid to proteins is called myristoylation. It is involved in key cellular functions including apoptosis, cellular differentiation and intracellular signalling. Myristic acid is vulnerable to redox damage and can be rendered immunogenic when oxidised [56].

In a previous study of patients with schizophrenia, Sethom observed a decrease in the levels of ω-6 (arachidonic acid) and ω-3 (DHA) PUFAs but an increase in the ω-6:ω-3 ratio [14]. The higher ω-6:ω-3 PUFA ratio may be indicative of a pro-inflammatory response [57,58]. However, a stronger inflammatory response may increase the production of free radicals and reduce PUFA levels [59]. Reduced anti-inflammatory activity may be implicated in the negative symptoms and cognitive impairment observed during the acute stages of schizophrenia episodes [60]. Our finding of a higher ω-6:ω-3 ratio related to negative symptoms is consistent with previous research.

Among ω-9 MUFAs, eicosenoic acid and erucic acid (its elongation product) was correlated with negative symptoms and were higher in the negative-symptoms-dominant group. Eicosenoic acid was recently reported as a key biomarker for schizophrenia in a large global metabolic profiling study [61]. Serum eicosenoic acid levels were significantly higher in patients with schizophrenia, prior to antipsychotic treatment than in healthy controls [61]. Elevated eicosenoic acid (together with other observed metabolic abnormalities) may reflect increased FA catabolism as an alternative energy source, resulting from insufficient mitochondrial glycolytic energy generation in the brain [62]. The present findings indicate that such metabolic abnormalities occur in UHR individuals prior to schizophrenia, particularly in relation to negative symptoms. Increased levels of eicosenoic and/or erucic acid have also been reported in the brains of rats with cerebral ischemia [63], in cognitively impaired elderly people [64], and in children with regressive autism [65]. These observations suggest that increases in ω-9 FAs may be associated with central nervous system dysfunction [66].

Negative symptoms generally begin prior to schizophrenia onset and are regarded as a more primary psychopathology [67]. Indeed, the disability of patients with schizophrenia is associated primarily with negative symptoms. The negative symptoms of schizophrenia are less responsive to antipsychotic treatment than the positive symptoms, and specific treatment for negative symptoms is not yet available. Thus, our findings of an association between membrane FAs and negative symptoms may help to improve our understanding of schizophrenia and aid in the development of treatments for UHR individuals or patients with schizophrenia. Further studies are needed to identify the specific types of symptoms that respond best to treatment with ω-3 FA supplements.

Several limitations should be considered when interpreting these data. First, the sample size was relatively small and there were a large number of fatty acids assessed; as such, these data need to be regarded as exploratory, as there is a risk of both type 1 and 2 errors. Since all consecutive referrals over a period of approximately 2 years were considered for inclusion the sample is representative for UHR patients [17]. It is not known if the associations between alterations in FAs and the severity of symptoms are attributed to primary factors (i.e. the neurobiology underlying schizophrenia spectrum disorders) or secondary factors (i.e. altered dietary patterns arising from altered disease-related behaviour). Use of a UHR cohort goes some way to reducing this confound.

Our findings suggest that abnormalities in FA metabolism may be associated with the neurobiological etiology of psychopathology in UHR individuals. Alterations in membrane FAs correlated with negative symptoms, which are primary psychopathological features of schizophrenia and are closely associated with the long-term outcome of schizophrenia spectrum disorder. Membrane FAs could serve as useful biological markers for early stages of schizophrenia; modulation of membrane FAs may provide a preventative intervention in subjects at UHR for schizophrenia.