Proteome analysis of schizophrenia patients Wernicke's area reveals an energy metabolism dysregulation

All samples were obtained from the brain bank of the Central Institute of Mental Health (Mannheim, Germany). Controls were collected at the Institute of Neuropathology, Heidelberg University, and clinical records were collected from relatives and general practitioners. Patient samples were derived from in-patients of the Mental State Hospital Wiesloch, Germany. All cases and controls were German whites. All SCZ patients were long-term in-patients at the Mental State Hospital Wiesloch, Germany, and SCZ diagnosis was made ante-mortem by an experienced psychiatrist according to the DSM-IV criteria [23]. For each patient the antipsychotic treatment history was assessed by examining the medical charts and calculated in chlorpromazine equivalents (CPE), through the algorithm developed by Jahn and Mussgay [24] for typical neuroleptics and clozapine. The CPE of olanzapine was calculated using the mean doses described by Meltzer and Fatemi [25]. All patients and controls underwent neuropathologic characterization to rule out associated neurovascular or neurodegenerative disorders. The classification according to Braak was stage II or less for all subjects [26, 27]. Patients and controls had no history of alcohol or drug abuse and severe physical illness. All assessments and post-mortem evaluations and procedures were previously approved by the ethics committee of the Faculty of Medicine of Heidelberg University, Germany. Detailed patient information is given in Additional file 1.

Fifty milligrams of human WA were individually homogenized in 1.5 ml tubes with glass spheres in 200 μl of 7 M urea, 2 M thiourea, 4% CHAPS, 2% ASB-14 and 70 mM DTT buffer [28]. Samples were centrifuged for 10 min at 14,000 rpm and quantified [29]. Supernatants were used for 2-DE.

Using two-dimensional gel electrophoresis we performed an individual proteome analysis of 9 schizophrenia samples compared to a pool of 6 healthy control. The experiments were carried out as described by Martins-de-Souza et al. [28]. Briefly, 650 μg of protein from SCZ or control samples were applied to IPG gel strips with a nonlinear separation range of pH 3–10 prior the second dimension in 12.5% T acrylamide gels that were stained using a colloidal coomassie blue protocol.

All experiments were performed in triplicate. Only proteins that appeared to be differentially expressed in triplicate gels were considered as differentially regulated.

2-DE gel images were used for spot detection and pI/MW calibration of WA samples using the ImageMaster 2D software, (GE Healthcare, Uppsala, Sweden). A total of thirty 2-DE profiles were analyzed, consisting of 27 gels from patients (9 patients evaluated individually in triplicates) plus three gels for the controls (triplicates of pooled controls). Volumes of all spots in SCZ and CTRL gels were determined and corresponding spots were matched for all 2-DE profiles. The spot volumes were analyzed by Kruskal-Wallis one-way analysis of variance aiming to reveal the statistically significant differences in the protein expression between the two groups. Protein spots with a mean n-fold change between SCZ and CTRL WA gels of +/- 1.8 were excised for mass spectrometry (MS) identification. False discovery rate (FDR) was calculated according to Storey, 2002 [30]http://​www.​genomine.​org/​qvalue/​index.​html. Protein identification by peptide mass fingerprinting was done as described by Martins-de-Souza et al. [31].

For western blot analysis, proteins were extracted as described above. One-hundred μg of total protein from dorsolateral prefrontal cortex (DLPFC) were run on a 12% SDS minigel (Bio-Rad, Hercules, CA, USA) and the proteins were transferred to Immobilon PVDF membranes (Millipore, Bedford, MA) at 100 V for 1 h with cooling. Membranes were treated with 5% Carnation instant nonfat dry milk in TBS-T for 4 hours, rinsed in TBS-T 3 times for 20 minutes and incubated with anti-PRDX6 (Abcam, Cambridge, UK) or anti-GFAP antibodies (Abcam, Cambridge, UK) at a 1:1000 dilution in TBS-T overnight at 4°C. In the next day, membranes were washed with water and TBS-T for 15 min. Incubations with anti-c-MYC-peroxidase antibody (GE Healthcare, Uppsala, Sweden) were carried out for 40 min at room temperature, followed by membrane washing with water and TBS-T and incubation with ECL mixture (GE Healthcare, Uppsala, Sweden) for 1 min and ECL film exposure (GE Healthcare, Uppsala, Sweden). Films were scanned, and the band signals (optical densities) were assessed using QuantityOne software (Bio-Rad, Hercules, CA, USA).

Differentially expressed proteins were classified according to their main biological processes as well as their molecular functions using Human Protein Reference Database (HPRD – http://​www.​hprd.​org). Moreover, the WA differentially expressed proteins were classified according their biochemical pathways using KEGG analysis http://​www.​genome.​jp/​kegg through software "Blast2Go" http://​www.​blast2go.​org.

Most proteins we identified as differentially expressed (52%) are enzymes involved the regulation of energy metabolism. Their altered levels can lead to an overall disturbance, and ultimately contribute to the establishment of pathological states [39, 40]. In this regard, a recent report has suggested that brain evolution from monkeys to humans may have led to the appearance of SCZ due to changes affecting energy pathway genes [41]. The alteration of this pathway, specifically the glycolysis and gluconeogenesis pathways in the WA is strongly supported by our data (Figure 2C).

Enzymes with functions in glycolysis such as ALDOC (Aldolase C), TPI1 (Triosephosphate isomerase 1), GAPDH (Glyceraldehyde-3-phosphate dehydrogenase), PGAM1 (Phosphoglycerate mutase 1) and ENO2 (Enolase 2) were found to be differentially expressed in WA, and are very likely to affect the glucose metabolism in SCZ patients. Likewise a dysfunction in ATP metabolism can result from differentially expressed ATP synthase subunits, ATP6V1A (ATPase, H+ transporting, lysosomal 70 kDa, V1 subunit A) and ATP5A1 (ATP synthase subunit alpha, mitochondrial). The above-mentioned proteins, and in some cases their genes, have been previously found to be differentially expressed in SCZ (Additional file 2).

We also found the differential regulation of dihydropteridine reductase (QDPR upregulated: 2.25×), which has been previously found to be differentially regulated [7] in 10 SCZ patients compared to 10 paired controls. This enzyme is a key in the control of dopamine and serotonin synthesis [42].

In our analyses we also found two new energy-metabolism-related proteins to be regulated in SCZ, which may have important roles SCZ negative symptoms: dihydrolipoyl dehydrogenase (DLD: upregulated: 2.17×) and peroxiredoxin 6 (PRDX6 upregulated: 2.33×). DLD is a component enzyme of complexes such as pyruvate dehydrogenase, alpha-ketoglutarate dehydrogenase, and the branched-chain alpha-keto acide dehydrogenase which participate in cell redox processes. PRDX6 is a bifunctional enzyme with two active sites with distinct roles. This enzyme is involved in redox regulation protecting the cell against oxidative injury by reducing levels of H₂O₂, short chain organic fatty-acids and phospholipid hydroperoxides [43]. Oxidative stress that can lead to DNA damage, protein inactivation, altered gene expression and apoptotic events, probably starts in SCZ during neurodevelopment. The close connection of these mechanisms to neuronal plasticity suggests that an oxidative component may be relevant to disease pathogenesis [39, 40, 44]. Thus, it is plausible that the differential regulation of an oxidative protector like PRDX6 may contribute to pathogenesis. The phospholipase A2 (PLA2) activity of PRDX6 on the other hand is critical for the regulation of phospholipid turnover [43]. An accelerated PLA2-mediated phospholipid turnover was previously observed in SCZ frontal lobe [45] whereas an increased activity of Ca²⁺-dependent PLA2 may lead to a reduced brain dopaminergic activity [45, 46]. The differential regulation of PRDX6 may accelerate the phospholipid turnover, leading to the described differential dopaminergic state. As both PRDX6 activities seem to be related to SCZ pathogenesis, we analyzed PRDX6 protein expression by western blot (WB) in the DLPFC of SCZ patients (Figure 3A). Again, a significant differential expression of PRDX6 was found (Figure 3B). Moreover, it is interesting to note that PRDX6 gene is localized in a chromosome loci (1q25.1) previously related with SCZ [47].

Although a correlation between the differential expression of glucose metabolism and oxidative phosphorylation enzymes has been recently described in SCZ [48], there is no consensus whether the protein expression alterations are causes or consequences of a pathological process. Also, the mitochondrial protein alterations found in SCZ could be consequences of the medications used by patients [32]. However, alterations in the glucose metabolism have been extensively reported as a central component of SCZ and are unlikely due to antipsychotic effects [41, 49].

Tubulins are involved in multiple activities, including mitosis, cytokinesis, and vesicular transport. Alterations of tubulin subunits have been described in SCZ [7, 12, 31, 50] and were also found in WA in our study (Additional file 2). Importantly, possible roles of tubulin in SCZ have been suggested by Denarier et al. [51] who generated a mouse SCZ model by knocking out the stable tubulin-only polypeptide (STOP) gene, which is responsible for microtubule stabilization in neurons. STOP null mice have a normal brain anatomy but show synaptic function defects due to dysfunctional glutamatergic transmission [52]. Moreover, this model exhibits a hyperdopaminergic state in the limbic system [53].