Introduction

Schizophrenia is a chronic mental disorder that usually appears at the end of adolescence or the beginning of adulthood and develops slowly for months or even years [1]. About 1 % of the world population has this condition, which presents a heritability of 80–85 % [2] and may reduce life expectancy by almost 20 years [3]. The many symptoms of schizophrenia are classified into positive symptoms, such as hallucinations, and thought disorders, and negative symptoms, such as lack of interest in social interaction, lack of motivation, and anhedonia. Cognitive deficiencies, such as the reduction in executive functions, selective attention, working memory, and mental flexibility, may also be present [4]. As a multifactorial disease, schizophrenia involves exogenous and endogenous factors as early as the beginning of neurodevelopment. Some molecular aspects of the pathology are still to be unraveled, and knowledge about the connections between the known aspects has to be improved for a more integrated understanding of schizophrenia physiopathology.

The corpus callosum (CC) is the largest portion of white matter in the human brain; it is located at its center between the right and left hemispheres and is responsible mainly for inter-hemispheric communication [5]. Morphological, electrophysiological, and neurophysiological studies have suggested alterations in the CC of schizophrenia patients [6–8]. These findings indicate the pivotal importance of the CC in establishing and maintaining schizophrenia, claiming for the understanding of its molecular features. Proteomics is a suitable tool for this purpose, since it can contribute to understanding biological and molecular processes through the integrated identification of unregulated biochemical pathways [9].

Our research group and others have contributed to this issue by analyzing the transcriptome and proteome of postmortem brain tissues and providing evidence on the potential role of oligodendrocytes and myelination in schizophrenia [10, 11]. These studies have also identified a large number of differentially expressed proteins associated with energy metabolism, the cytoskeleton, and cell signaling [12]. Since these pathways are mainly in the cytoplasm, the present study used an enrichment method to obtain only the soluble fraction of the CC proteome and pinpoints exactly differences in the expression of glial proteins, because the CC is predominantly a white matter region. The method used here is suitable for studies when one needs to obtain a detailed coverage of the proteome of interest [13].

In this paper, we describe our efforts in deciphering the cytoplasmic proteome of the CC in schizophrenia and mentally healthy controls. We chose the CC because of its importance in schizophrenia and also because of evidence on the role of glia cells in the disease. Although this brain region has been previously studied [14], the methodology used here is the state of the art in the field [15]; in contrast, in their study in 2007, Sivagnanasundaram et al. employed two-dimensional gel electrophoresis (2DE), which has intrinsic limitations for large-scale proteome analysis [16]. We cross-validated our results by comparing them with those obtained in other brain regions.

Materials and methods

Human samples

The CC samples (with no definition if from the anterior, middle, or posterior part) were provided by the BrainNet Europe consortium. Samples were collected postmortem from nine chronic schizophrenia patients with residual symptoms (diagnosed antemortem by an experienced psychiatrist according to the DSM-IV criteria) and seven controls (Table 1). Patients’ samples came from the State Mental Hospital, Wiesloch, Germany, while control samples came from the Institute of Neuropathology, Heidelberg University, Heidelberg, Germany. The controls had not had any kind of brain disorder or somatic disease and had not taken any antidepressant or antipsychotic medications. Each schizophrenia patient had a record of antipsychotic treatment, so we calculated chlorpromazine equivalents (CPE). The CPE for typical neuroleptics and clozapine were calculated with Jahn and Mussgay’s algorithm [17], while the CPE for olanzapine was calculated according to Meltzer and Fatemi [18]. Patients and controls were German Caucasians with no history of alcohol or drug abuse. Brains were submitted to neuropathological characterization to rule out any associated brain disorder. The brains analyzed here presented Braak staging less than II. All assessments, postmortem evaluations, and procedures were approved by the ethics committee of the Faculty of Medicine, Heidelberg University, Heidelberg, Germany.

Table 1 Clinical data of patients and controls
Full size table

Sample preparation

Cytosolic proteins were obtained from the brain tissues according to the protocol developed by Cox and Emili [19]. Twenty milligrams of each CC sample was homogenized in ten volumes of 0.32 M sucrose (Sigma-Aldrich, St. Louis, MO, USA) and 4 mM HEPES (Sigma-Aldrich) buffer (pH = 7.4), and one tablet of protease cocktail inhibitor was added (Roche Diagnostics, Indianapolis, IN, USA) per 25 ml of buffer. The homogenate was centrifuged at 1000×g for 10 min at 4 °C. Pellets were dissolved in ammonium bicarbonate 50 mM prior to protein digestion.

Nano-liquid chromatography-mass spectrometry (nano LC–MS/MS) analyses

Each sample was digested in solution overnight with a 1:80 trypsin:total protein ratio. Next, the resulting peptides were lyophilized and frozen prior to mass spectrometric analyses. Just before analysis, peptides were dissolved in 0.1 % formic acid aqueous solution and injected into a nano-LC system comprising an autosampler and 2D-nano high-performance liquid chromatography (HPLC; Eksigent, Dublin, CA, USA), coupled online to an LTQ-Orbitrap XL mass spectrometer (Thermo Scientific, Bremen, Germany). The full detailed description of the nano LC–MS/MS configuration and data analyses can be found in Maccarrone et al. [20].

Proteome quantification

CC cytosolic proteomes were quantified by the label-free spectral counting approach using Mascot Distiller (Matrix Sciences, London, UK). Mascot Distiller aligned the mass spectrometric data from each CC cytosolic sample by using mass and elution time. The quantification was based on the relative intensities of extracted ion chromatograms (XICs). FDR was set to a minimum of 0.1 and based in at least five independent MS/MS spectra.

Results and discussion

We identified a total of 5678 unique peptides that corresponded to 1636 proteins belonging to 1512 protein families on the CC cytosolic proteome. Of these proteins, 65 were differentially expressed in patients with schizophrenia compared with mentally healthy controls: 28 were upregulated and 37 downregulated (Table 2).

Table 2 Proteins differentially expressed in schizophrenia corpus callosum, classified according to their biological and molecular functions
Full size table

Of the proteins differentially expressed in schizophrenia patients, 31 % are related to cell growth and maintenance, 28 % to energy metabolism, and 20 % to cell communication and signaling. Among the proteins related to cell growth and maintenance, nine are structural, nine are cytoskeletal, and two are from the extracellular matrix; among those related to energy metabolism, 17 are metabolic enzymes; and among those related to cell communication and signaling, six are adapter molecules (Table 2), which are essential for the pathways in which they are involved. The enzymes altered in energy metabolism are key molecules to regulate the energy metabolism pathway [21, 22], and the structural and cytoskeleton proteins are indispensable to maintain cellular shape [23–26], in addition to playing a role in information transmission from dendrites to axons [27], a requirement for the cell to properly exercise its function. The receptor molecules at the membranes are related to cell communication, and, in the case of neurons, they are essential for a proper synaptic function [28, 29].

As can be seen in Fig. 1, Ingenuity Pathway Analysis (IPA, Ingenuity Systems, Qiagen, Redwood, CA, USA; www.ingenuity.com) also showed that these proteins clustered within those pathways. IPA is based on an algorithm that uses curated connectivity information from the literature to determine the network of interactions among the differentially expressed proteins and canonical pathways in which they are involved [30].

Fig. 1
figure 1

Network of interactions among differentially expressed proteins according to an analysis of biological systems by Ingenuity Pathway Analysis

Full size image

As shown in Fig. 1, three predominant pathways have differentially expressed proteins in schizophrenia: energy metabolism, cell communication and signaling, and cell growth and maintenance. These pathways have already been found to be dysregulated in patients with schizophrenia [14, 23, 31]. This protein network also evidences connections among differentially expressed proteins within their biological processes and between different processes, suggesting dysfunctional networks of proteins, as is known to be the case in schizophrenia [32].

In addition, information provided about altered canonical pathways in schizophrenia samples shows that differently expressed proteins are related to each other as part of common pathways, such as cell metabolism and cell signaling, as shown in Fig. 2, further reinforcing those pathways’ dysregulation, as shown in Fig. 1. All pathways were significantly altered in the schizophrenia samples compared to the control samples (p < 0.05). The information also indicates that canonical pathways are associated to oxidative stress such as superoxide radical degradation and NRF2-mediated oxidative stress response.

Fig. 2
figure 2

Canonical pathways associated with differentially expressed proteins by Ingenuity Pathway Analysis

Full size image

We compared our results with those obtained by Sivagnanasundaram et al. [25]. Using the combination of 2DE and mass spectrometry (2DE-MS), they identified 34 differentially expressed proteins in the CC in schizophrenia, nine of which overlap with the data we present here. To validate our findings and also to highlight proteins consistently different in expression in schizophrenia patients’ brains, we compared our results to 16 other brain tissue studies of the proteome of schizophrenia. Thirty of the proteins differentially expressed in our study overlap with these studies (Table 3).

Table 3 Proteins consistently differentially expressed across different brain regions as referenced below
Full size table

Several studies consistently found differential expression of proteins associated with 14-3-3-mediated signaling, the pathway in which we found the main disruption in the CC of schizophrenia patients (Fig. 2). 14-3-3 protein zeta/delta (YWHAZ), 14-3-3 protein epsilon (YWHAE), and 14-3-3 protein gamma (YWHAG) are associated with cell communication and signaling, as binding proteins to several protein kinases and phosphatases, and also are involved in actin dynamics. They were found dysregulated also in other reports on the schizophrenia proteome [32]. The 14-3-3 protein family is highly associated with neurotransmitting processes [28], with some isoforms particularly enriched in synapses, and the YWHAZ and YWHAE genes were previously associated with schizophrenia [33–35]. For instance, broad 14-3-3 functional knockout mice, in addition to showing schizophrenia-like behavior, have high dopamine levels and a decrease in dendritic complexity and spine density, defects linked to schizophrenia [36]. As the main white matter region in the brain, the CC is composed mostly of glia cells and neuronal axons. Because these axons are responsible for the connection between the two brain hemispheres, the dysregulation of proteins involved in cell communication and neurotransmission may be pivotal to impairments in brain connectivity, as previously suggested [6–8].

Genome-wide association studies (GWAS) found that the gene for the protein clathrin heavy chain 1 (CLTC) is associated with schizophrenia. The protein, that has been found downregulated in our results (Table 2), has the function of a vesicle coat and plays a role in clathrin-mediated endocytosis and thus is involved in cell signaling [37]. According to Schubert el al. [38], clathrin influences processes such as synaptic dysfunction, white matter changes, and aberrant neurodevelopment and consequently may be related to schizophrenia pathology.

We found differential expression of the light and medium neurofilaments (NEFL and NEFM), as did Sivagnanasundaram et al. [25]. Differential expression of NEFM was found also by five other studies and differential expression of NEFL by two, highlighting their importance in schizophrenia. These proteins, which are coded by gene regions that are usually altered in schizophrenia patients [23, 26], are associated with cell growth and maintenance, reinforcing the role of this pathway in schizophrenia [23, 25, 39]. Neurofilaments play also important roles in the function of oligodendrocytes, cells enriched in brain white matter. Along these lines, we observed also the downregulation of myelin basic protein (MBP) and myelin–oligodendrocyte glycoprotein (MOG), the main constituents of myelin sheaths. These proteins were previously found differentially expressed in other schizophrenia brain regions [10, 32, 39], a finding that is in line with evidence at the transcriptome [40, 41] and morphological levels [42, 43]. The differential expression of MBP and MOG supports the idea that a degenerative process might be occurring in schizophrenia brains [10, 44–46]. Also, MBP and MOG might be biomarker candidates to be further investigated in combination with other potential biomarkers, since they were found in different concentrations in the cerebrospinal fluid (CSF) of living schizophrenia patients [23].

The family of tubulin proteins also was associated with the cell cytoskeleton. The isoform TUBB was found in our study and another ten studies (Table 3), which reinforces its role in schizophrenia pathobiology. According to Moehle et al. [47], tubulin proteins are the major protein components in axons, dendrites, and dendritic spines, and knockdown studies indicate that the different isoforms have distinct roles in neurodevelopment. Studies with mice knocked out for the STOP gene (stable tubulin-only polypeptide) have associated the family of tubulin proteins with synaptic dysfunctions probably caused by glutamatergic transmission dysregulation [14, 48].

The protein gelsolin (GSN) belongs to a superfamily of actin-binding proteins (ABPs) [49] and is related to actin remodeling on cell growth and apoptosis [50]. Our studies and that of Prabakaran et al. [51] found this protein to be downregulated, and transcriptome studies found its gene to be altered [10]. Considering that this protein is related also to myelin sheath generation and maintenance [44], it may be involved in myelination dysfunction in schizophrenia patients.

Another protein found upregulated here and differentially expressed in other four schizophrenia proteome analyses was gamma-enolase (ENO2). This enzyme participates in the glycolytic pathway, a pathway found to be the second most affected in this study (Fig. 2) and also by others [22]. ENO2 interconverts 2-phosphoglycerate to phosphoenolpyruvate [21]. Since glia cells are responsible for supplying energy to neuronal axons [52], this glycolysis dysregulation may affect negatively synaptic activity and neuronal plasticity [25, 39, 53]. Another three glycolytic enzymes were differentially expressed in our study but not in the study by Sivagnanasundaram et al. [25]. These three are pyruvate kinase (PKM), fructose-bisphosphate aldolase A (ALDOA), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), confirming the dysregulation in schizophrenia of glycolysis, a pivotal energy pathway [22].

In agreement with Sivagnanasundaram et al. [25] and another two studies (Table 3), we found differential expression of superoxide dismutase [Cu–Zn] (SOD1), which has been observed consistently to be dysregulated in the brain and liver of patients with schizophrenia [22, 54–57]. Our study and another two studies (Table 3) also found another protein related to oxidative stress, peroxiredoxin-1 (PRDX1), to be downregulated, reinforcing the concept of a reduced response to oxidative stress [58, 59].

Conclusions

Given the central role of the corpus callosum, our data revisited proteome alterations of this brain region with a more powerful proteomic methodology. To provide further evidence for the dysregulation of signaling and structural proteins, we investigated the cytosolic fraction of the CC proteome. We found not only these pathways dysregulated, but also we were able to reinforce the role of energy metabolism and oligodendrocytes in schizophrenia, despite the data provided in this study should further be validated by other orthogonal methods (i.e., western blot, immunocytochemistry). Our findings increase the understanding of schizophrenia pathophysiology and could be targeted for further treatment strategies.