Impaired Membrane Lipid Homeostasis in Schizophrenia

Abstract

Background and Hypothesis

Multiple lines of clinical, biochemical, and genetic evidence suggest that disturbances of membrane lipids and their metabolism are probably involved in the etiology of schizophrenia (SCZ). Lipids in the membrane are essential to neural development and brain function, however, their role in SCZ remains largely unexplored.

Study Design

Here we investigated the lipidome of the erythrocyte membrane of 80 patients with SCZ and 40 healthy controls using ultra-performance liquid chromatography-mass spectrometry. Based on the membrane lipids profiling, we explored the potential mechanism of membrane phospholipids metabolism.

Study Results

By comparing 812 quantified lipids, we found that in SCZ, membrane phosphatidylcholines and phosphatidylethanolamines, especially the plasmalogen, were significantly decreased. In addition, the total polyunsaturated fatty acids (PUFAs) in the membrane of SCZ were significantly reduced, resulting in a decrease in membrane fluidity. The accumulation of membrane oxidized lipids and the level of peripheral lipid peroxides increased, suggesting an elevated level of oxidative stress in SCZ. Further study of membrane-phospholipid-remodeling genes showed that activation of PLA2s and LPCATs expression in patients, supporting the imbalance of unsaturated and saturated fatty acyl remodeling in phospholipids of SCZ patients.

Conclusions

Our results suggest that the mechanism of impaired membrane lipid homeostasis is related to the activated phospholipid remodeling caused by excessive oxidative stress in SCZ. Disordered membrane lipids found in this study may reflect the membrane dysfunction in the central nervous system and impact neurotransmitter transmission in patients with SCZ, providing new evidence for the membrane lipids hypothesis of SCZ.

Introduction

Schizophrenia (SCZ) is a serious mental disorder affected by genetic and environmental risk factors, which is one of the top ten causes of disability.¹ However, the etiology and pathophysiology of SCZ have not been fully elucidated. SCZ is characterized by a variety of systemic abnormalities, such as abnormalities in neurodevelopment,² inflammatory responses,³ and energy metabolism.⁴ Therefore, studies on the etiology of SCZ must consider the systemic nature of the disorder. Years ago, there were studies showing that the phospholipid dissociation process of cell membranes was abnormal in patients with SCZ.^{5, 6} Subsequently, Horrobin^{7, 8} proposed the membrane phospholipid hypothesis, suggesting that SCZ is a disorder of membrane lipid metabolism, and this abnormality of membrane lipid affects neurological abnormalities and complex brain behaviors. This hypothesis partially explains the characteristics observed in SCZ from genetic and biochemical perspectives.

In the past decade, mass spectrometry (MS) and bioinformatics software have developed rapidly in the field of lipidomics. These technological developments are helpful for understanding the role of lipids in health and disease and for shedding light on the potential functions of lipids. Notably, there are approximately 100 000 different lipids in humans, which contribute greatly to maintain the normal structure and function of cells.⁹ As integral components of cellular membranes, lipids directly control the biophysical parameters of membrane; for example, the relative size and degree of saturation of fatty acyl in lipids affect membrane curvature, fluidity, and thickness.¹⁰ Fluidity is an important factor in physicochemical properties of the membrane¹¹ in vivo as it directly affects cell membrane fusion, material transport, and signal transduction. In addition, lipids can modulate the activity of membrane proteins with lipid-binding domains by recruiting them to specific membrane compartments or subdomains.¹² Membrane lipid homeostasis was affected by a variety of functional abnormalities, including those caused by exposure to an environment of higher oxidative stress¹³ and dysregulation of lipid biosynthesis and catabolic pathways.¹⁴ In the central nervous systems (CNS), membrane lipid disorders affect the functions of synapses¹⁵, release of neurotransmitter transmissions¹⁶, and second messengers¹⁷, such as lysophospholipids (LPLs), diacylglycerol (DAG), and eicosanoids. In brain, membrane-derived signaling lipids influences subjective perception, mood, and emotional behaviors.¹⁸ However, the roles of lipids and lipid-like molecules in SCZ are largely unknown and need to be further explored.

Lipidomics of peripheral^{19, 20} and brain tissues²¹ has suggested that aberrant lipidome composition changes occur in psychiatric disorders. In earlier studies, SCZ patients had significantly decreased phosphatidylethanolamine (PE) and increased sphingomyelin (SM) in the red blood cell (RBC) membrane.^{22, 23} Membrane phosphatidylcholine (PC) is significantly altered in different brain regions of SCZ patients.²⁴ In addition, changes in lipids already occur in the high-risk clinical stages of psychosis, suggesting that abnormal lipid metabolism may be an important feature of SCZ.^{25, 26} Subtle aberrant changes in lipids or their fatty acyl chains have a noticeable effect on membrane biological functions, and understanding these alterations is promised to reveal the biochemical basis of dysfunction in SCZ.²⁷ Due to the limitations of objective factors, including lipid extraction methods, MS platforms, and lipid databases, the results of existing studies are inconsistent and need to be further improved. In detail, first, only small amounts of lipids have been reported in membrane of SCZ. Second, early studies mainly reported the total content of membrane phospholipid species, but few studies involved the level of individual lipids. Third, few studies have reported the levels of fatty acyl chains from different lipids in the membrane in SCZ. Finally, the regulatory mechanism of membrane lipids in SCZ needs to be further explored. Therefore, a comprehensive and systematic investigation of membrane lipids in SCZ was the main objective of this study.

The erythrocyte membrane has multifaceted advantages for the study of lipidomes. As we mentioned before, lipids play important physiological and pathological roles in both periphery and CNS²³, and lipid signaling in the brain is linked to the composition of RBC membrane lipids, such as the polyunsaturated fatty acids (PUFAs) content in RBC were positively correlated with the levels of phospholipid in brain.^28–32 In addition, the lipids in the erythrocyte membrane are less affected by short-term diet than circulating triglycerides and reflect long-term fatty acid (FA) intake, which is the cumulative result of the interaction of genetics, metabolism, and dietary intake.³³ The lipids in the RBC membrane are more stable than those in plasma or serum.³⁴ Thus, in this study, we investigated the lipid profiling of RBC membrane in SCZ to identify the membrane lipids associated with this disorder. Furthermore, we attempted to gain insights into the regulatory mechanism of membrane lipid homeostasis under the pathological state of SCZ, and further deepened the membrane lipid hypothesis of SCZ.

Methods

In this study, we examined the erythrocyte membrane lipidome in SCZ and HCs, and further explored the mechanism of membrane lipid metabolism by detecting plasma oxidative stress markers and expression levels of phospholipid remodeling gene in leukocyte. Details of the methods were described in the supplementary materials.

Results

Abnormal Membrane Lipid Size and Diversity in Schizophrenia

The global lipid content in erythrocyte membranes was detected in 80 patients with SCZ and 40 HCs using UHPLC-MS (supplementary table S1). In supplementary figure S1, we show the range of lipids contained in the erythrocyte membrane based on m/z and retention time. A large group of glycerophospholipids (PLs) were detected in both the nonpolar positive and nonpolar negative, while glycerides were mainly detected in the nonpolar-positive dataset. Among 945 lipids, 812 lipids were quantified, including PEs, PCs, phosphatidylserines (PSs), phosphatidylinositols (PIs), SMs, ceramides (CERs), triacylglycerols (TAGs), DAGs, lysophosphatidylethanolamines (LPEs), lysophosphatidylcholines (LPCs), and lysophosphatidylserines (LPSs), a total of 11 lipid classes (figure 1). Multivariate statistical analysis of the 812 lipids showed significant differences between SCZ patients and HCs (R² = 0.947, Q² = 0.755) (figure 2A).

Fig. 1.

Predominant lipid classes in RBC membrane. Characterizing the number and diversity of lipids detected in erythrocyte membrane based on classifications in the LipidSearch database. Percentage (%) indicates the proportion of the number of lipids in different classes to the total lipids.

Fig. 2.

Membrane lipids were obviously different between SCZ patients and HCs. (A) The PLS-DA model was used to distinguish RBC membrane lipids between in SCZ patients and in HCs. (B) Percentage of the 11 major lipid classes within the RBC membrane between SCZ patients and HCs (Percentage: mean ± SD; unit: %). (C) Differences in 812 lipids between SCZ patients and HCs (The q-value corrected by FDR).

The analysis of the percentage of the 11 lipid classes showed a higher percentage of PI (FC = 1.05, P = 6.85E-03) and lower percentages of LPC (FC = 0.81, P = 2.70E-05) and LPE (FC = 0.92, P = 3.99E-02) in patients with SCZ compared to HCs (figure 2B and supplementary table S2). After multiple comparison correction, 387 lipids were found to be differentially abundant (q < 0.05) between the SCZ and HCs (supplementary table S3). Among these lipids, 244 lipids were significantly reduced in SCZ, and 154 of the decreased phospholipids were categorized as PEs or PCs (figure 2C). Ether phospholipids consist of two types of phospholipids with an ether (alkyl–acyl phospholipid) or vinyl ether (alkenyl–acyl phospholipid, also known as plasmalogen) bond at the sn-1 position of the glycerol backbone. The decrease in individual alkyl–acyl phospholipids observed in SCZ patients was marked (supplementary figure S2A), and the total concentration of alkyl-acyl phospholipids was significantly decreased in both the PC and PE classes (FC = 0.83, P = 2.95E-10 and FC = 0.93, P = 7.88E-08, supplementary figure S2B and S2C). Among the ether lipids, the total amount of plasmalogen was significantly reduced in SCZ (FC = 0.91, P = 3.51E-06). Furthermore, we found an independent biomarker, PE(O-18:3_20:4), an ether phospholipid, that can effectively distinguish patients with SCZ from the HCs (AUC = 0.97, sensitivity = 0.938, specificity = 0.950) (supplementary figure S3).

The Composition of Saturated and Unsaturated Fatty Acyl Chains of Membrane Lipids was Imbalanced in SCZ

We found the high content of RBC membrane FAs, including C16:0, C18:0, C18:1, C18:2. C20:4, and C22:6, were mainly from PC, PE, and PS. While the long-chain FAs, such as C24:0, C24:1, and C24:2, were mainly derived from SM and CER (supplementary figure S4). The abundances of different FAs in the erythrocyte membrane were obviously different; for example, the saturated fatty acids (SFAs) were mainly C16:0 and C18:0, the MUFAs with the highest content were C18:1, and the PUFAs with high contents were C20:4 and C18:2 (figure 3). We analyzed the contents of 58 FAs by adding up the contents of the same fatty acyl chain in the 11 lipid classes (supplementary table S4). The results showed that C16:0 with higher abundance in SFAs, was significantly increased in CER, PC, PE, PI, and DAG. In contrast, most of the PUFAs in PE were decreased, especially C18:2, C20:4, C22:3, and C22:4, and the total concentrations of MUFAs and PUFAs in the membrane were significantly reduced in SCZ (figure 3A and supplementary table S5). The global variation in these individual FAs affect the corresponding indicators, reflecting the status of the erythrocyte membrane. Among membrane FA-indexes., saturation indexes (SI) (% SFA/% MUFA, P = 2.34E-07 and % SFA/% PUFA, P = 1.97E-14) were significantly increased in SCZ (figure 3B and 3C). The lipophilic index was significantly higher in the SCZ patients (P = 7.41E-09, figure 3D), reflecting the lower membrane fluidity in patients with SCZ. Other significantly reduced indicators of SCZ include the unsaturation index (UI) (P = .045, figure 3E) and chain length index (CLI) (P = 3.79E-04, figure 3F). There was no significant difference in the peroxidation index (PI) between SCZ patients and HCs (P = .627, figure 3G).

Fig. 3.

The level of unsaturated fatty acids in the RBC membrane of SCZ patients was decreased, and FA indices differed in SCZ. (A) Membrane lipidome results showed that acyl chains of 11 lipid classes, described by the number of double bonds and chain length. The direction of the fold change and significance by heat map. The bubble size represents the mean of different fatty acyl chain concentration percentages in the same lipid classes in HCs. SFA, MUFA, and PUFA represent the total content of three types of fatty acyl chains in different lipid classes (The q-values were corrected by FDR, q < 0.05). Data in (B-G) represent FA indexes of the membrane, the calculation method please see the additional methods in Supplementary Materials. (B-D) Membrane fluidity indexes: total SFA%/MUFA%, total SFA%/PUFA%, and the lipophilic index. (E). The unsaturation index. (F) The chain length index. (G) The peroxidation index, showing FA susceptibility to peroxidation by calculating the double bond content. The p-values were corrected by two-tailed T-test (P < .05).

Increased Oxidative Stress in Schizophrenia

We identified 67 oxidative lipids and found that 44.8% of oxidized lipids were significantly increased in SCZ (figure 4A and supplementary table S6). The total concentration of plasmalogen ethanolamine (plas-PE) was significantly lower in patients with SCZ (figure 4B). Based on these results, to determine the redox state in SCZ patients, we measured the plasma levels of 4-hydroxynonenoic acid (4-HNE), a kind of lipid peroxide. Our results showed that the plasma 4-HNE level was higher in patients with SCZ than in HCs (FC = 1.49, P = 6.11E-05, figure 4C). Moreover, superoxide dismutase (SOD), an antioxidant enzyme, can scavenge free radicals, which total SOD activity in the plasma was significantly lower in SCZ than in HCs (FC = 0.84, P = .012, figure 4D).

Fig. 4.

Impaired membrane lipid homeostasis was related to the increase of oxidative stress and abnormal phospholipid remodeling. (A) Significant changes in oxidized lipids of erythrocyte membrane in SCZ patients (The p-values corrected by Mann-Whitney U-test and the q-values were corrected by FDR, details are in supplementary table S6). (B) The total amount of plasmalogen ethanolamine in SCZ patients was significantly reduced. (FC = 0.91, P-value = 3.51E-06). (C) Plasma lipid peroxides (4-HNE) were significantly elevated in SCZ (FC = 1.49, P = 6.11E-05). (D) The plasma activity of SOD was significantly decreased in SCZ (FC = 0.84, P = .012). (E) The mRNA levels of LPCATs and PLA2s were detected in patients with SCZ (n = 80) and HCs (n = 60). The p-values were calculated based on logistic regression correction for age, BMI and sex. One asterisk indicates the P-value < .05, two asterisks indicate the P-value < .01, three asterisks indicate the P-value < .001, and four asterisks indicate the P-value < .0001.

Activated Expression of Enzymes Involved in Membrane Phospholipid Remodeling in Patients with SCZ

We found that the lipids ratio (LPC/PC) in the membrane of SCZ patients was significantly lower than that of HCs (FC = 0.83, P = .0001, supplementary figure S5). This reflected the abnormal activity of proteins related to membrane lipid remodeling. We assessed the mRNA expression levels of members of the lysophosphatidylcholine acyltransferases (LPCATs) and the phospholipase A2 (PLA2) family (PLA2G4A and PLA2G6) involve in membrane phospholipid remodeling. There was a significant increase in LPCAT1 (FC = 1.19, P = .043) and LPCAT4 (FC = 1.49, P = 1.09E-05) in patients with SCZ after adjusting for age, sex, and BMI. Similarly, the mRNA levels of PLA2G4A and PLA2G6 in the SCZ group were increased (FC = 2.17, P = 1.35E-06 and FC = 1.18, P = .020) (figure 4E and supplementary table S7). In contrast, no significant changes in LPCAT2 and LPCAT3 were observed between the SCZ patients and HCs. We further established a model of oxidation-damaged neuron cells (PC12 differentiated cells) and found that two subtypes of PLA2s (PLA2G4A and PLA2G6) and LPCAT4 in the LPCATs family were activated, consistent with the results observed in SCZ patients (supplementary figures S6–S8).

Discussion

In this study, disordered lipid homeostasis was observed in the RBC membrane in SCZ, mainly manifested as the significantly decreased contents of individual PCs, PEs, and PUFAs. Moreover, we explored the mechanism of abnormal membrane lipids in SCZ. Our study demonstrated that SCZ patients have elevated oxidative stress, and influence the expression of membrane phospholipid remodeling enzymes, showing activation of PLA2s and LPCATs. These results indicated that the dysregulation of membrane lipid and the related biochemical processes were potential pathological mechanisms of SCZ (figure 5).

Fig. 5.

Related mechanisms and pathology of membrane lipid homeostasis disorder in SCZ. Membrane lipid homeostasis was impaired in patients with schizophrenia. Excessive oxidative stress resulted in membrane lipid peroxidation. During the remodeling process of membrane phospholipids, the expression of PLA2s and LPCATs is activated, resulting in an imbalance of SFA and PUFA composition in membrane phospholipids, which further leads to the decrease of membrane fluidity. Disturbance of the membrane-lipid system is pervasive throughout the body of SCZ patients, leading to membrane dysfunction (such as abnormal neurotransmitter transmission between synapses), and subsequent development of schizophrenia. PLs, glycerophospholipids; LPLs, lysophospholipids; OxPLs, oxidized phospholipids; OxFAs, oxidized fatty acids; PLA2s, iPLA2, and cPLA2; LPCATs, LPCAT1, and LPCAT4; 4-HNE, 4-hydroxynonenonic acid; ROS, reactive oxygen species.

Potential Mechanism of Impaired Membrane Lipid Homeostasis in Patients with SCZ

There are two possible mechanisms leading to the disruption of membrane lipid homeostasis. On the one hand, previous studies have shown that decreased levels of membrane phospholipid precursors in the brains of patients with SCZ suggest reduced de novo synthesis of PCs and PEs.³⁵ On the other hand, our results provide some evidence to support that impaired membrane lipid homeostasis in patients with SCZ related to the imbalanced decomposition and remodeling of phospholipids under the elevated oxidative stress.

First, the levels of oxidative stress were elevated in patients with SCZ compared with HCs. Plasmalogen is a kind of endogenous antioxidant, and its vinyl ether constitutes an important target of ROS.³⁶ In our study, total plas-PE and SOD activities were significantly reduced in SCZ, suggesting that the antioxidant capacity of patients with SCZ was weakened. We found a large accumulation of oxidized lipids in the erythrocyte membrane and elevated plasma peroxides in patients with SCZ. Under high oxidative stress, the reaction of ROS with the double bonds of unsaturated acyl chains in membrane phospholipids produces oxidized lipids³⁷, such as arachidonic acid (AA), which dissociated from the membrane to form 4-HNE and malondialdehyde through nonenzymatic oxidation reaction.³⁸ Taken together, these results suggest that elevated oxidative stress in SCZ affected the instability of phospholipid and disrupts lipid homeostasis in the membrane.

Second, the decomposition and incorporation of fatty acyl chains of membrane phospholipids in SCZ were imbalanced. The diversity and saturation of fatty acyl chains in each phospholipid need to be modified by a phospholipid deacylation and reacylation pathway mediated by the PLA2s and lysophopholipid acyltransferases (LPLATs) referred to as the Lands cycle.³⁹ We found that the mRNA levels of PLA2G4A and PLA2G6, genes that encode calcium-dependent phospholipase A2 (cPLA2) and calcium-independent phospholipase A2 (iPLA2), respectively, were significantly increased in SCZ patients. Similarly, a series of studies have shown that the activities of PLA2s were increased in the peripheral and brain tissues of SCZ patients.^40–42 Elevated oxidative stress in SCZ patients leads to membrane lipid peroxidation, and recruitment of PLA2s for excision of the oxidized PUFAs.⁴³ Furthermore, cPLA2 preferentially hydrolyses AA at the sn-2 position of phospholipids to generate lysophospholipids, and iPLA2 shows a slight preference for AA over other FAs.⁴⁴ Activated PLA2s lead to the excessive decomposition of membrane phospholipids, accompanied by a reduction in PUFAs, especially AA, which was consistent with our results.

Third, another important finding in our study is that the expression of LPCATs was upregulated in SCZ patients. Remarkably, the members of the LPCAT family exhibit different acyl-CoA preferences. LPCAT1 prefers palmitoyl-CoA (16:0-CoA) as the acyl donor to produce dipalmitoyl PC.⁴⁵ The activity of LPCAT2 is the highest in the presence of arachidonoyl-CoA (20:4-CoA).⁴⁶ The preferred substrate for LPCAT3 is polyunsaturated fatty acyl CoA (18:2-CoA or 20:4-CoA), while the preferred substrate for LPCAT4 is oleoyl-CoA (18:1-CoA).⁴⁷ The expression of LPCAT1 and LPCAT4 was significantly upregulated in patients with SCZ. Among PC species, dipalmitoyl PC and C16:0 were significantly increased, while the linoleoyl and arachidonoyl chains of PC were significantly reduced in SCZ. Dysregulation of LPCATs affects the selectivity of FAs incorporated into membrane phospholipids in SCZ. In other words, PUFA dissociation and SFA incorporation in membrane phospholipids were enhanced in SCZ patients. It is reasonable to presume that abnormal expression of enzymes related to phospholipid remodeling is one of the reasons for the imbalance of phospholipid saturation in the membrane in patients with SCZ.

Pathological Implications of Impaired Membrane Lipid Homeostasis in SCZ

We found that the significantly decreased individual PEs and PCs in SCZ patients were mainly ether phospholipids. Recently, studies have also found reduced serum levels of ether phospholipids in patients with SCZ.^{19, 48} Ether lipids are a primary component of myelin in the CNS,⁴⁹ and vesicles or membranes containing ether phospholipids contribute to fusion and reduce surface tension, which play important roles in physiological processes that depend on exocytosis or endocytosis.⁵⁰GNPAT knockout mice lacking ether lipids, exhibit an imbalance in synaptic vesicle circulation, decreased neurotransmitter release, hyperactivity, and impaired social interaction.⁵¹ Yao et al. found that peripheral and CNS membrane phospholipid turnover was tightly coupled.³¹ It is conceivable that membrane lipid homeostasis is impaired throughout the body in SCZ.⁵² In fact, nerve cells develop in the early stage of life and have a long-life span, which makes them more vulnerable to the accumulation of damage from the internal and external environment. Membrane lipids are ubiquitous in synapses of the brain and contain a high proportion of PUFAs vulnerable to oxidative stress.¹³ In addition, consistent with our results, the levels of PCs in thalamic membrane were reduced in SCZ patients.⁵³ Peripheral membrane lipids of SCZ may be attacked by ROS and disturbed by abnormal regulation of synthesis and metabolism, leading to impaired membrane lipid homeostasis. We hypothesized that damage to peripheral membrane lipid also occurred in the CNS and affected the pathological development of SCZ.

In our results, the total PUFAs and AA in the erythrocyte membrane in patients with SCZ were significantly decreased. Notably, AA was significantly decreased in postmortem brain tissues from SCZ patients, which is in line with the results we found in the RBC membrane.⁵⁴ Alterations in membrane PUFAs affect the level of neurotransmitters (e.g., dopamine, serotonin (5-HT), and glutamate) and downstream signaling transduction,⁵⁵ such as AA mediated the regulation of endocannabinoid, which inhibit the release of neurotransmitters (including glutamate, gamma-aminobutyric acid, and acetylcholine).⁵⁶ As important second messengers, PUFAs, which are released from or incorporated into the membrane, reflect the activity of neuroreceptors.⁵⁷ In addition, excessive dissociation of membrane PUFAs leads to the development of inflammation cascades, which has attracted attention in the study of the molecular pathology of SCZ.

The imbalance of SFAs and PUFAs in the membrane, observed in patients with SCZ, can influence membrane fluidity. An earlier study using fluorescent polarization spectroscopy showed that erythrocyte membrane fluidity is reduced in patients with SCZ.⁵⁸ Perturbations in membrane fluidity influence the binding sites on receptors and/or the density of the receptors in the membrane,⁵⁹ especially in the postsynaptic membrane. In the presynaptic terminals, which are efficient membrane-remodeling machines, lipids regulate membrane trafficking and cooperate with proteins underlies presynaptic activity.⁶⁰ PUFAs reduce membrane bending stiffness and contribute to cellular membrane deformation processes; for example, supplementation of exogenous AA in Caenorhabditis elegans can regulate synaptic vesicle recycling⁶¹ and long-term feeding of AA in rats contributed to maintain membrane fluidity in neuron.⁶² Several lines have shown that the density and affinity of 5-HT receptors were lower in patients with SCZ than in HCs and this alteration was affected by the reduction of membrane PUFAs.⁶³ Overall, we hypothesized that dysregulation of membrane lipid homeostasis leads to systemic membrane dysfunction in patients with SCZ, including an impact on the synaptic membrane system, which further affect neurotransmitter systems.

There are some limitations in this study. Although we detected the RBC membrane lipids in participants from the same area (similar diet habit) and their BMI were carefully matched between SCZ and HC groups, it will be more rigorous to include an assessment of food intake to adjust for its effect on membrane lipid levels. In addition, because most of the patients in this study were under medication, antipsychotic effect on lipid metabolism was evaluated. Considering that the effect of antipsychotics on RBC membrane lipids was inconclusive in previous studies, and there was no correlation between antipsychotics and membrane lipids in SCZ patients in our study (data not shown), the antipsychotics effect on membrane lipids were quite limited. Admittedly, the results of this study need to be further verified in larger scale, multi-center, and multi-geographic populations before clinical application.

This study reported a high-throughput and high-resolution lipidomic profiling of erythrocyte membranes in SCZ, which enabled us to gain a comprehensive understanding of membrane lipid homeostasis in SCZ patients. Furthermore, we reported the abnormal expressions of LPCATs, which could be potential targets for the treatment of abnormal membrane phospholipid turnover in SCZ. The abnormal phospholipid remodeling and impaired membrane lipid homeostasis we observed in SCZ patients was possibly caused by enhanced oxidative stress. This study provides new insights for the underlying pathological mechanisms of SCZ, facilitating clinical diagnosis and intervention, and provides a basis for adjuvant therapies in SCZ patients, such as supplementation of PUFAs to regulate membrane lipids in a long-term longitudinal cohort in the future study.

Funding

This work was supported by the National Natural Science Foundation of China (Grant No. 81971254, Grant No. 81771440, Grant No. 81901354), the National Key Research and Development Program of China (Grant No. 2016YFC1306900, Grant No. 2016YFC1306802), Shanghai Municipal Science and Technology Major Project (Grant No. 2017SHZDZX01), Interdisciplinary Program of Shanghai Jiao Tong University (Grant No. YG2019GD04, Grant No. ZH2018ZDA40), and the Foundation of Shanghai Key Laboratory of Psychotic Disorders (Grant No. 13DZ2260500), Furthermore, this study was supported by China Postdoctoral Science Foundation (Grant No. 2019M661526, Grant No. 2020T130407), Natural Science Foundation of Shanghai (Grant No. 20ZR1426700), and Youth Research Initial Foundation of Shanghai Jiao Tong University (Grant No. 19X100040033).

Competing interests

All authors declare no conflict of interest.