Essential polyunsaturated fatty acids and social cognition in schizophrenia

1. Introduction 

Disturbances of cognitive functions, such as memory, executive function, verbal fluency and attention, are characteristic of most patients with schizophrenia or related disorders (Saykin et al., 1991, Sumiyoshi et al., 2001a, Sumiyoshi et al., 2001b, Sumiyoshi et al., 2004, Sumiyoshi et al., 2005, Matsui et al., 2004, Keefe et al., 2005). Cognitive performance has been shown to have a significant influence on important outcome measures, including work and social function. Among tests of cognitive performance, script tasks have been developed as a measure of social cognition (Chan et al., 1999, Matsui et al., 2006). These tasks are used to measure the ability of subjects to evaluate component actions of social situations. Generally, a script task consists of free recall, frequency judgment, and sequencing of events in the schema of daily activities, such as dining at restaurants or shopping at supermarkets (Chan et al., 1999, Matsui et al., 2006). One of the most consistent findings with the script tasks has been that patients with schizophrenia perform worse on judging the events that sometimes happen (middle frequency events), rather than those that usually or rarely happen (prominent events), in the Frequency Judgment Task (Chan et al., 1999, Matsui et al., 2006).

Altered composition of phospholipids, a major component of neural membranes, has been suggested to be related to the pathophysiology of schizophrenia (Horrobin, 1998). Among the constituents of phospholipids, saturated fatty acids, e.g. palmitic acid (14:0, PA) and stearic acid (16:0, SA), as well as monounsaturated fatty acids, e.g. oleic acid (18:1 n-9, OA) and nervonic acid (24:1, NA), can be synthesized de novo. On the other hand, essential polyunsaturated fatty acids (EPUFAs) must be ingested in the diet. These include: linoleic acid (18:2 n-6, LA), dihomogammalinolenic acid (20:3 n-6, DGLA), and arachidonic acid (20:4 n-6, AA) as the n-6 series; and eicosapentaenoic acid (20:5 n-3, EPA), docosapentaenoic acid (22:5 n-3, DPA), and docosahexaenoic acid (22:6n n-3, DHA) as the n-3 series (Horrobin, 1998, Fenton et al., 2000).

Since EPUFA levels in erythrocyte membranes have been shown to reflect EPUFA composition in the brain (Connor et al., 1990), various studies have been conducted to evaluate the change in this peripheral measure of EPUFAs in patients with schizophrenia (for review, see Fenton et al., 2000). These studies indicate an overall decrease in the concentrations of the above-mentioned EPUFAs in patients compared with normal control subjects (Vaddadi et al., 1989, Glen et al., 1994, Yao and van Kammen, 1994, Peet et al., 1995, Khan et al., 2002, Arvindakshan et al., 2003a, Arvindakshan et al., 2003b). Diet, medication status, or smoking may not be associated with alterations in the EPUFA levels (Yao and van Kammen, 1994, Yao et al., 1994a, Yao et al., 1994b, Doris et al., 1998, Assies et al., 2001, Reddy et al., 2004). On the other hand, erythrocyte membrane levels of EPUFAs, such as AA, have been shown to be correlated with the severity of psychotic symptoms (Peet and Horrobin, 2002, but see Assies et al., 2001).

In addition to the relationship between abnormal EPUFA metabolism and schizophrenia, previous studies also suggest altered fatty acid compositions in patients with depression (Maes et al., 1996, Maes et al., 1999) or autism (Vancassel et al., 2001), as well as clinical benefits from the administration of EPUFAs to subjects with borderline personality disorder (Zanarini and Frankenburg, 2003) or antisocial behavior (Gesch et al., 2002). Since these disorders are characterized by disturbances of social abilities, it is hypothesized that decreased EPUFA levels may be associated with impaired performance on neuropsychological tests measuring social cognition.

So far, there has been little study on the relationship between abnormal EPUFA metabolism and disturbances of cognitive performance relevant to social functions in subjects with schizophrenia. The primary purpose of the present study, therefore, was to determine whether decreased erythrocyte membrane EPUFA levels would be correlated selectively with poor performance on script tasks, but not with other aspects of cognitive dysfunctions, in patients with schizophrenia. We also sought to determine if EPUFA concentrations are related to the severity of psychotic symptoms.

2. Methods 

2.1. Subjects 

Twenty-five chronically-ill out-patients (male/female=14/11) meeting DSM-IV criteria for schizophrenia entered the study. They were recruited from the Outpatient Clinic of the University Hospital of Toyama. All available clinical information and data were obtained from a structured clinical interview (Sumiyoshi et al., 2001a, Sumiyoshi et al., 2001b). Subjects were diagnosed by a consensus of at least two experienced psychiatrists. Patients known to be abusing alcohol or other illicit drugs, or those with epilepsy, brain damage, or neurologic disorders, were excluded from the study. This study was carried out in accordance with the latest version of the Declaration of Helsinki. Written informed consent was obtained after the explanation of the study. The protocol was approved by the Institutional Review Board of the University of Toyama School of Medicine. Sixteen patients were treated with the following first generation (typical) antipsychotic drugs: haloperidol (N=8), nemonapride (N=4), zotepine (N=2), chlorpromazine (N=1), and levomepromazine (N=1); nine patients were neuroleptic-free. The mean (S.D.) daily haloperidol-equivalent dose (mg) for the patients on medications was 6.9 (7.3). Patients were assessed with the Brief Psychiatric Rating Scale (BPRS) (Overall and Gorham, 1962) by psychiatrists or well-trained clinical psychologists. Interrater reliability was >0.80 (Sumiyoshi et al., 2001a, Sumiyoshi et al., 2001b). Blood samples were obtained from these subjects between 09:30 and 10:00 h after an overnight fast.

Blood samples were also obtained from 32 age- and gender-matched healthy volunteers (male/female=18/14) recruited from the hospital staff, university students, and volunteers from the community. Subjects were excluded if they had a history of psychiatric illness, head trauma, neurological illness, serious medical or surgical illness, or substance abuse. None of the control subjects were receiving pharmacological treatment for medical illnesses. Candidates were excluded if they had any personal or family history of psychiatric illness. The healthy control group was also screened for a history of psychiatric disorders in their first-degree relatives. The demographic data of the subjects are summarized in Table 1.

Table 1. Demographic data of subjects

		Control	Schizophrenia
	N	32	25
	% Male (N)	56.3 (18)	56.0 (14)
	Age, years	33.0 (8.0)	34.6 (12.3)
	Age of onset, years	–	25.1 (7.2)
	BPRS score
	Total	–	17.6 (10.9)
	Positive	–	6.4 (5.1)
	Negative	–	4.6 (3.6)

Values represent mean (S.D.). BPRS, Brief Psychiatric Rating Scale.

2.4. Analysis of fatty acids 

Erythrocyte membrane fatty acid levels were analyzed based on an established method using a gas chromatography system (Ranjekar et al., 2003). Briefly, 1 ml of RBC obtained from subjects was collected into a 15-ml screw cap vial that subsequently received 4.0 ml of 0.6 N methanolic HCl containing 4 μl of 0.5% BHT. The vials were sealed and incubated at 80 °C for 2 h. After incubation, methylated fatty acids were extracted twice with hexane. Layers were separated by centrifugation in a swinging rotor at 3000×g for 15 min at room temperature. The hexane layers were carefully removed and collected in a separate vial. The hexane extract was completely dried by passing argon and stored at −40 °C until use. The methylated fatty acids were resuspended in 150 μl hexane. Aliquots (1 μl) were used for the fatty acid analysis with a Shimadzu gas chromatograph, Model GC-2010 (Japan), using a capillary column of 30 m×0.32 mm×0.20 μm in dimensions (Supelco, USA). A flame ionization detector was used with a column oven temperature at 160 °C for 10 min, programmed at 10 °C rise/min up to 175 °C and finally held at 220 °C for 10 min. The temperature of the injector and detector was set at 240 °C and 275 °C, respectively. The column was calibrated by injecting the standard fatty acid mixture in approximately equal proportions. The peaks in recorded data were identified as per the retention time of the standard fatty acids run under the identical conditions.

Fatty acid data were grouped into saturated fatty acids (PA, SA), monounsaturated fatty acids (OA, NA), and EPUFAs [n-6 (LA, DGLA, AA), n-3 (EPA, DPA, DHA)]. The levels of these fatty acids are reported as the percent of the total fatty acids (Khan et al., 2002) (Table 2).

Table 2. Fatty acid compositions

		Controls	Treated patients	Drug-free patients
	Saturated
	PA	22.7 (0.8)	26.1 (5.1)	23.3 (0.8)
	SA	19.7 (0.7)	22.4 (3.8)	20.6 (1.8)
	Monounsaturated
	OA	13.6 (0.7)	15.1 (1.8)	13.8 (1.0)
	NA	1.3 (0.1)	1.3 (1.3)	1.0 (0.4)
	Polyunsaturated
	n-3 series
	EPA	1.5 (0.7)	1.5 (1.1)	1.8 (1.7)
	DPA	3.3 (0.5)	2.8 (1.5)	3.4 (1.0)
	DHA	9.5 (1.2)	7.4 (3.4)	9.1 (1.5)
	Subtotal	14.3 (1.8)	10.8 (6.1)	14.3 (3.0)
	n-6 series
	LA	11.1 (1.2)	9.3 (1.2)	9.7 (1.3)
	DGLA	1.6 (0.2)	1.5 (0.6)	1.6 (0.3)
	AA	15.7 (1.6)	12.5 (5.1)	15.7 (2.2)
	Subtotal	28.4 (2.1)	21.7 (8.6)	27.0 (2.8)

Values represent mean (S.D.) of the percent of the total fatty acids (see Section 2). PA, palmitic acid; SA, stearic acid; OA, oleic acid; NA, nervonic acid; EPA, eicosapentaenoic acid; DPA, docosapentaenoic acid; DHA, docosahexaenoic acid; LA, linoleic acid; DGLA, dihomogammalinolenic acid; AA, arachidonic acid.

3. Results 

Age did not differ between control subjects and patients with schizophrenia (t=0.62, n.s.). MANOVA of the fatty acid measures revealed a significant overall effect between normal controls and patients with schizophrenia (Wilks' lambda=0.80, df=3,53, P<0.01). Subsequent analysis showed that EPUFA levels (F=9.10, df=1,55, P<0.01), but not saturated (F=0.40, df=1,55, n.s.) or monosaturated (F=0.04, df=1,55, n.s.) fatty acid levels, were significantly lower in the patients than in the normal controls (Table 2 and Fig. 1). None of the clinical measures were significantly different between neuroleptic-free subjects and those receiving antipsychotic drugs (data not shown).

View Large Image Download to PowerPoint

Fig. 1. Erythrocyte membrane fatty acid levels. EPUFAs, essential polyunsaturated fatty acids. ⁎P<0.01, significantly lower compared with normal controls.

Since the above results suggest a selective change in EPUFA levels in the patients, the correlations between this fatty acid measure and scores on the Frequency Judgment Task, Digit Span, AVLT, and BPRS subscales were analyzed. The numbers of correctly answered items [mean (S.D.)] for the High, Middle, and Low Frequency Events in the Frequency Judgment Task in the patients were 6.2 (2.5), 3.0 (1.1), and 3.3 (0.8), respectively. The patients showed normal estimated IQ [103.0 (12.1)]. A positive correlation was found between EPUFA levels and the score for judgment of the Middle Frequency Events (r=0.47, P<0.05), but not that of the High (r=0.04, n.s.) or Low (r=0.03, n.s.) Frequency Events. There was no significant correlation between the Middle Frequency Events Judgment score and IQ (r=0.12, n.s.). Also, the BPRS Positive score was negatively correlated with EPUFA levels (r=−0.42, P<0.05), while no significant correlations were found for the Total (r=−0.22, n.s.) or Negative (r=0.05, n.s.) subscale score. Performance on the Digit Span [9.9 (2.1)] or AVLT [31.3 (7.5)] was not correlated with EPUFA concentrations (r=−0.10, n.s. and r=0.10, n.s., respectively).

4. Discussion 

The results of this study confirm previous reports indicating abnormal metabolism of EPUFAs in schizophrenia. Moreover, the present study provides, to our knowledge, the first suggestion for a link between decreased levels of EPUFAs and impaired social cognition, as represented by event schema, in schizophrenia. Also, the decrease in EPUFA levels was positively correlated with the severity of positive symptoms.

Decreased erythrocyte membrane levels of individual EPUFAs, such as LA, AA, and DHA, have been reported in subjects with schizophrenia (Vaddadi et al., 1989, Glen et al., 1994, Yao and van Kammen, 1994, Peet et al., 1995, Khan et al., 2002, Arvindakshan et al., 2003a, Arvindakshan et al., 2003b, Reddy et al., 2004). The two saturated (PA, SA) and two monounsaturated (OA, NA) fatty acids, together with the six EPUFAs (LA, DGLA, AA, EPA, DPA, DHA), reported here, constitute more than 90% of the total fatty acids in erythrocytes and also in the brain (Khan et al., 2002). Therefore, the observed decrease in the EPUFA levels in subjects with schizophrenia is consistent with the previous studies reporting altered composition of phospholipids in the illness. Specifically, the selective change in polyunsaturated, but not saturated fatty acid levels, in subjects with schizophrenia (Fig. 1) is in line with previous studies (e.g. Assies et al., 2001, Reddy et al., 2004).

Data presented in the current study indicate the relationship between decreased EPUFA compositions and the severity of impaired performance on frequency judgment of the script tasks in subjects with schizophrenia. Several lines of recent research have focused on functional capacity of patients with schizophrenia, i.e. an individual's capacity for performing key tasks of daily living (Green et al., 2004). The script tasks have been considered to provide a test of functional capacity, or proxy measures of social problem solving and daily activities (Matsui et al., 2006). Impaired learning performance has been found in n-3 fatty acid-deficient rats (Wainwright, 1992, Fenton et al., 2000 for review), which may be associated with cognitive changes consistent with decreased function of the dopaminergic activity in the prefrontal cortex (Fenton et al., 2000). Specifically, erythrocyte membrane EPUFA levels have been shown to reflect EPUFA composition in the frontal cortex in rhesus monkeys (Connor et al., 1990). On the other hand, performance on the script tasks has been demonstrated to be related to frontal lobe activity (Crozier et al., 1999). Taken together, it is possible that the altered phospholipids composition in the frontal cortex, as represented by decreased EPUFA levels in erythrocyte membranes, may be a biological basis for poor performance on the script tasks in subjects with schizophrenia. Although the lack of a correlation between EPUFA concentrations and attention or verbal memory, reported here, deserves further scrutiny, the present findings may provide a rationale for future studies of the effect of treatment with EPUFAs on social and other domains of cognition in schizophrenia. The positive correlation between the severity of positive, but not negative symptoms, and the degree of the decrease in EPUFA levels are partly in line with the findings of Assies et al. (2001).

The limitations of the current study should be noted here. First, the results of this study were based on a relatively small sample size, which is prone to a type-I and/or type-II error. Second, this study lacks more detailed information about the subjects, such as dietary habits and smoking, although these variables may not be a significant confounding factor (Assies et al., 2001, Reddy et al., 2004). Third, although the cross-sectional reductions of EPUFA levels were observed in patients with schizophrenia, the main findings from the present study are correlational rather than causational. Further investigations on the relationship between fatty acid metabolism and various domains of cognitive function are warranted.