PLOS ONE RESEARCH ARTICLE Calcium-binding proteins are altered in the cerebellum in schizophrenia Francisco Vidal-Domènech1,2☯, Gemma Riquelme1☯, Raquel Pinacho1☯, Ricard Rodriguez-Mias3, América Vera1, Alfonso Monje4, Isidre Ferrer5, Luis F. Callado6,7,8, J. Javier Meana6,7,8, Judit Villén3, Belén Ramos ID1,2,4,7* 1 Psiquiatria Molecular, Institut de Recerca Sant Joan de Déu, Esplugues de Llobregat, Spain, 2 Dept. de Bioquı́mica i Biologia Molecular, Facultat de Medicina, Universitat Autònoma de Barcelona, Bellaterra, Spain, 3 Department of Genome Sciences, School of Medicine, University of Washington, Seattle, Washington, United States of America, 4 Parc Sanitari Sant Joan de Déu, Sant Boi de Llobregat, Spain, 5 Departamento a1111111111 de Patologia y Terapeutica Experimental, Universidad de Barcelona, Senior consultant Servicio Anatomia a1111111111 Patológica, Hospital Universitario de Bellvitge-IDIBELL, CIBERNED, Hospital de Llobregat, Barcelona, Spain, a1111111111 6 Department of Pharmacology, University of the Basque Country UPV/EHU, Leioa, Bizkaia, Spain, 7 Centro a1111111111 de Investigación Biomédica en Red de Salud Mental, Madrid, CIBERSAM, Spain, 8 Biocruces Bizkaia Health Research Institute, Barakaldo, Spain a1111111111 ☯ These authors contributed equally to this work. * bramos@fsjd.org OPEN ACCESS Abstract Citation: Vidal-Domènech F, Riquelme G, Pinacho R, Rodriguez-Mias R, Vera A, Monje A, et al. Alterations in the cortico-cerebellar-thalamic-cortical circuit might underlie the diversity of (2020) Calcium-binding proteins are altered in the symptoms in schizophrenia. However, molecular changes in cerebellar neuronal circuits, cerebellum in schizophrenia. PLoS ONE 15(7): e0230400. https://doi.org/10.1371/journal. part of this network, have not yet been fully determined. Using LC-MS/MS, we screened pone.0230400 altered candidates in pooled grey matter of cerebellum from schizophrenia subjects who Editor: Tomasz Boczek, Medical University of Lodz, committed suicide (n = 4) and healthy individuals (n = 4). Further validation by immunoblot- POLAND ting of three selected candidates was performed in two cohorts comprising schizophrenia Received: February 24, 2020 (n = 20), non-schizophrenia suicide (n = 6) and healthy controls (n = 21). We found 99 signif- icantly altered proteins, 31 of them previously reported in other brain areas by proteomic Accepted: June 17, 2020 studies. Transport function was the most enriched category, while cell communication was Published: July 8, 2020 the most prevalent function. For validation, we selected the vacuolar proton pump subunit 1 Copyright: © 2020 Vidal-Domènech et al. This is an (VPP1), from transport, and two EF-hand calcium-binding proteins, calmodulin and parval- open access article distributed under the terms of bumin, from cell communication. All candidates showed significant changes in schizophre- the Creative Commons Attribution License, which permits unrestricted use, distribution, and nia (n = 7) compared to controls (n = 7). VPP1 was altered in the non-schizophrenia suicide reproduction in any medium, provided the original group and increased levels of parvalbumin were linked to antipsychotics. Further validation author and source are credited. in an independent cohort of non-suicidal chronic schizophrenia subjects (n = 13) and non- Data Availability Statement: The mass psychiatric controls (n = 14) showed that parvalbumin was increased, while calmodulin spectrometry proteomics data have been deposited was decreased in schizophrenia. Our findings provide evidence of calcium-binding protein in the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier dysregulation in the cerebellum in schizophrenia, suggesting an impact on normal calcium- PXD008216. All relevant data are within the dependent synaptic functioning of cerebellar circuits. Our study also links VPP1 to suicide manuscript and its Supporting Information files. behaviours, suggesting a possible impairment in vesicle neurotransmitter refilling and Funding: This work was supported by a Miguel release in these phenotypes. Servet grant (MS16/00153-CP16/00153) to BR financed and integrated into the National R + D + I and funded by the Instituto de Salud Carlos III (Spanish Ministry of Health) – General Branch PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 1 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia Evaluation and Promotion of Health Research – Introduction and the European Regional Development Fund (ERDF), the Basque Government (IT1211/19) and Schizophrenia constitutes a complex disorder with a mixture of symptoms and cognitive defi- SAF 2017-88126-R to J.J.M. The funders had no ciencies. It is considered a brain disorder in which a malfunction of multiple regions in distrib- role in study design, data collection and analysis, uted brain macro- and microcircuits could lead to the different symptoms. According to the decision to publish, or preparation of the hypothesis of cognitive dysmetria, which is based on neuroimaging findings, the alteration of manuscript. specific components of the cortico-cerebellar-thalamic-cortical circuit could contribute to the Competing interests: The authors have declared impairment of mental coordination processes, leading to the emergence of symptoms in schizo- that no competing interests exist. phrenia [1–3]. The cerebellum, as part of this circuit, has been suggested to play a role in the production of the range of symptoms and cognitive impairments in schizophrenia. Much evi- dence has been collected over the last few decades supporting the involvement of the cerebellum in higher cognitive functions and in the pathophysiology of this neurodevelopmental disorder [4–6]. Based on the clinical signs of individuals with cerebellar lesions and functional neuroim- aging studies in patients with schizophrenia, it has been suggested that the impairment of the cerebellar function could affect many altered domains in schizophrenia, including executive functions, working memory, language, attention, social cognition and emotion (reviewed in [4,6,7]). The lateral cerebellar cortex is involved in some of these cognitive abilities, such as working memory, executive functions and speech [5,8–10]. The complex cytoarchitecture of the cerebellar cortex comprises a variety of neurons including GABAergic neurons, such as Pur- kinje cells, basket cells, stellate cells, and Golgi cells, and glutamatergic cells, such as granule neu- rons and unipolar brush cells. Both Purkinje cells and granule cells are key in orchestrating the organisation of this circuitry during development, and thus defective regulation of the intracel- lular mechanisms maintaining the normal synaptic functioning of these cells could impact on internal cerebellar circuit activities and the cerebellum output signal to the cortex. These altered cerebellar circuits have been linked to a number of neurological and psychiatric conditions such as schizophrenia (reviewed in [11,12]). Moreover, molecular hypothesis-driven, studies have found genes from the GABAergic and glutamatergic neurotransmission system with altered expression in the cerebellar cortex, supporting the idea that the regulation of these systems is also disrupted in the cerebellum [13–18]. Gene expression changes have been detected in the cerebellum using transcriptomic screenings in a few studies in schizophrenia [19–21]. However, to the best of our knowledge, proteomic studies to identify altered proteins in cerebellar neuro- nal circuits have not yet been performed. The cerebellar cortex is a highly homogeneous brain area composed mainly of cerebellar granule neurons. This characteristic makes the cerebellar cortex an attractive area for investigating proteomic changes in a relatively small sample. Here, we conducted an exploratory quantitative proteomic screen using pooled grey matter of the cerebellar cortex from schizophrenia (1 pool of 4 samples) and control healthy individu- als (1 pool of 4 samples) with the aim of identifying consistently altered proteins in schizophre- nia (SZ) (S1A Fig). We then validated three selected candidates by immunoblot in a total of 47 subjects, which included the same samples from the proteomic analysis, an extended cohort (control, n = 7; SZ, n = 7) with a non-schizophrenia suicide group (n = 6), and a larger inde- pendent cohort with non-suicidal chronic schizophrenia subjects (control, n = 14; SZ, n = 13) (S1A Fig). Our validation focused on three selected candidates with a possible relevant role in the disorder and detected with robust changes in the screening. Materials and methods Brain tissue samples For the proteomic analysis, we used post-mortem human brain tissue obtained from the UPV/ EHU brain collection, from the cerebellum of subjects with paranoid schizophrenia who had PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 2 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia committed suicide (n = 4) and control subjects who had died in a traffic accident and who had had no history of psychiatric episodes (n = 4) (Table 1 and S1 Table). Samples were obtained at autopsies in the Basque Institute of Legal Medicine, Bilbao, Spain, in compliance with the poli- cies of the research and ethical boards for post-mortem studies. Toxicological screening for antipsychotics, antidepressants, and other drugs was performed at the National Institute of Toxicology, Madrid, Spain. All deaths were subjected to retrospective analysis for previous medical diagnosis. Subjects with ante-mortem criteria for paranoid schizophrenia according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-IIIR and DSM-IV) were matched to control subjects who had died from accidental causes in a paired design, based on gender, age, and post-mortem delay (PMD). To validate the candidates identified in the quantitative proteomic assay, we also used an extended cohort from the UPV/EHU brain collection for individual sample analysis of human post-mortem cerebellum (Table 1, Cohort I). A total of 20 brains of subjects who had commit- ted suicide (n = 12), died in an accident (n = 5) or from natural causes (n = 3) were selected. Subjects with paranoid schizophrenia (n = 7; 6 suicide victims (falling from a height (n = 4) and hanging (n = 2)) and 1 non-suicide subject that died from natural causes), control subjects (n = 7) that had died in a traffic accident (n = 5) or from natural causes (n = 2), and a non- schizophrenia suicide victim group (n = 6; falling from a height (n = 5) and hanging (n = 1)) were matched by gender, age, post-mortem delay and pH (Table 1, Cohort I). Control subjects were chosen among the collected brains on the basis, whenever possible, of the following crite- ria: (a) negative medical information on the presence of neuropsychiatric disorders or drug abuse, (b) accidental or natural cause of death, (c) negative results in toxicological screening for psychotropic drugs except for ethanol, and (d) a post-mortem delay not longer than 48 hours. Diagnoses were established according to the DSM-IIIR or DSM-IV. Diagnoses in the non-schizophrenia suicide victim group included obsessive compulsive disorder (n = 1), depression (n = 2), anxiety disorder (n = 1), depression and personality disorder (n = 1), and alcohol dependence (n = 1). Samples were coded by the brain collection staff to protect human subject confidentiality. The study was approved by the Institutional Ethics Committee of the Fundació Sant Joan de Déu. For further validation analysis of protein candidates, we used a larger independent cohort of post-mortem human cerebellum of subjects with chronic schizophrenia (n = 13) who had died from natural causes and control subjects with no history of psychiatric episodes (n = 14) from the collection of neurologic tissues of Parc Sanitari Sant Joan de Déu [22] and the Insti- tute of Neuropathology Brain Bank (HUB-ICO-IDIBELL Biobank) (Table 1, Cohort II). These collections follow the guidelines of Spanish legislation and the approval of the local ethics com- mittees. Written informed consent was obtained from each subject. The study was approved by the Institutional Ethics Committee of Parc Sanitari Sant Joan de Déu. We matched schizo- phrenia and control groups by gender, age, post-mortem delay and pH. Table 1 (Cohort II) shows the demographic, clinical and tissue-related characteristics of the samples. All SZ sub- jects were institutionalized donors with a long duration of the illness (Table 1, Cohort II) who had no history of neurological episodes. Experienced clinical examiners interviewed each donor ante-mortem to confirm schizophrenia diagnosis according to DSM-IV and Interna- tional Classification of Diseases 10 (ICD-10) criteria. The last mean daily chlorpromazine equivalent dose for the antipsychotic treatment of patients was based on the electronic records of last drug prescriptions administered up to death (Table 1, Cohort II) and was calculated as previously described [23]. The mean and standard deviation of body mass index in patients was 24.22 ± 4.26. Possible tardive dyskinesia side effect of treatments was assessed in donors using the Abnormal Involuntary Movement Scale (AIMS) [24]. The total score was calculated using the sum of items 1 to 7, which assess the severity of abnormal movements in different PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 3 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia Table 1. Demographic, clinical and tissue-related features of cases of Cohorts I and II. Cohort I: Pilot Validation analysis (n = 20) Schizophreniaa (n = 7) Non-schizophrenia suicide (n = 6) Healthy Control (n = 7) SZ-C SZ-Suicide Statistic p-value Statistic p-value b b Gender N/A 1.000 N/A 1.000 Female 14% (n = 1) 17% (n = 1) 14% (n = 1) Male 86% (n = 6) 83% (n = 5) 86% (n = 6) c c Age (years) 44 ± 11 46 ± 11 45 ± 11 22.00 0.782 21.00 1.000 c c PMD (hours) 9.00 ± 3.37 9.00 ± 4.36 11.33 ± 8.36 21.00 0.683 19.50 0.858 c c pH 6.66 ± 0.60 6.90 ± 0.59 6.53 ± 0.42 19.50 0.556 19.50 0.882 Toxicology N/A N/A N/A N/A Atypical AP 42.8% (n = 3) 16.7% (n = 1) N/A Other 28.6% (n = 2) 50.0% (n = 3) 57.1% (n = 4) Drug-free 28.6% (n = 2) 33.3% (n = 2) 42.8% (n = 3) Somatic Disorders N/A N/A N/A N/A Diabetes 14.3% (n = 1) 14.3% (n = 1) Arterial hypertension 14.3% (n = 1) 14.3% (n = 1) Other 16.7% (n = 1) 14.3% (n = 1) Cohort II: Extended Validation cohort (n = 27) Non-suicide Schizophrenia (n = 13) Healthy Control (n = 14) Statistic p-value Gender Male 100% (n = 13) 100% (n = 14) N/A N/A Age (years) 72 ± 9 70 ± 11 0.72; 25d 0.478 PMD (hours) 5.62 ± 2.34 5.46 ± 1.82 0.02; 25d 0.984 pH 6.88 ± 0.49 6.61 ± 0.63 1.25; 25d 0.223 SZ diagnosis N/A N/A N/A Chronic residual 69.2% (n = 9) Chronic paranoid 15.4% (n = 2) Chronic disorganized 7.7% (n = 1) Chronic catatonic 7.7% (n = 1) Age of onset of SZ (years) 22 ± 8 N/A N/A N/A Duration of illness (years) 51 ± 9 N/A N/A N/A Daily AP dose (mg/day)e 562.92 ± 514.13 N/A N/A N/A Atypical AP 7.7% (n = 1) N/A N/A N/A Typical AP 92.3% (n = 12) N/A N/A N/A Somatic Disorders Diabetes 7.7% (n = 1) 28.6% (n = 4) N/A 0.326b Arterial hypertension 15.4% (n = 2) 57.1% (n = 8) N/A 0.046b Dyslipidemia 7.7% (n = 1) 35.7% (n = 5) N/A 0.165b Others 15.4% (n = 2) 7.1% (n = 1) N/A N/A Mean ± standard deviation or relative frequency are shown for each variable; PMD, post-mortem delay between death and brain sample collection; SZ, schizophrenia; C, Healthy control group; AP, antipsychotics; N/A, not applicable. a Paranoid schizophrenia (n = 7). b Frequencies were analysed using Fisher’s exact test. c Mann-Whitney U is shown for non-parametric variables. d T-statistic and degrees of freedom are shown for parametric variables. e Last chlorpromazine equivalent dose was calculated based on the electronic records of drug prescription of the patients. https://doi.org/10.1371/journal.pone.0230400.t001 PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 4 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia regions of the body. Each item was classified from 0 to 4 according to the severity (0-absence, 1-minimum, 2-mild, 3-moderate, 4-severe). The mean and standard deviation of the AIMS total score was 4.17 ± 5.22 (n = 12) with a minimum value of 0 and a maximum value of 15. 38.5% (n = 5) of the subjects showed no symptoms. 23.1% (n = 3) of the subjects showed minimum severity of symptoms, 23.1% (n = 3) mild and 7.7% (n = 1) moderate. No subjects showed severe symptoms. Samples were coded by each brain bank or collection staff to protect human subject confidentiality. Protein extraction Specimens of the lateral cerebellar cortex from the posterior lobe, extending from the pial surface to white matter and only including grey matter, were dissected from coronal slabs stored at -80 ˚C. Protein extracts were prepared from tissue samples using NP40 lysis buffer as described previously [25]. Protein concentration was determined by Bradford assay (Biorad). Mass spectrometry screening and data processing Our screening strategy combined differential isotopic labelling of peptides via reductive dimethylation with offline fractionation by SCX and liquid chromatography coupled to tan- dem mass spectrometry (LC-MS/MS) on a hybrid linear ion trap orbitrap mass spectrometer (LTQ-Orbitrap)) (S1A Fig). 400 μg of total protein extracts from four pooled control (100 μg/each) and four pooled schizophrenia (100 μg/each) lysates were digested and further processed as indicated in the S1 File. Briefly, the digested peptides were dimethyl-labelled with either hydrogen (light peptides, control) or deuterium (heavy peptides, schizophrenia) isotopes through a reductive dimethylation reaction as described previously [26]. Differen- tially labelled peptides were then mixed 1:1. Dimethylated peptide mixtures were separated by strong cation exchange (SCX) chromatography on a polysulphoethyl A column. Peptide mixtures were analysed by LC-MS/MS. Each peptide fraction was separated by reverse phase chromatography on a capillary column and analysed online on a hybrid linear ion trap orbi- trap (LTQ-Orbitrap XL, Thermo Scientific) mass spectrometer for identification and relative quantification of isotopically labelled peptide pairs. MS/MS spectra were searched against a concatenated target-decoy Uniprot human protein database (UP000005640 version 05-23- 2017, n = 71,567 target sequences) using the Comet search algorithm (version 2015025) and specific search parameters (see S1 File). The mass spectrometry proteomics data have been deposited in the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD008216 [27]. Peptide matches were filtered to <1% False-Discovery Rate (FDR) and protein groups were filtered at �90% probability score. The log2 heavy/light ratio for each protein was determined and transformed to a z-score [28]. A significance value (p-value) for each protein ratio was calculated from the complementary error function for the normalized distribution of the z-scores [28]. The FDR was computed for all the p-values using the Benjamini and Hochberg method [29]. The FDR threshold was set at 0.1 for selected significant proteins with consistent changes amongst peptides. Proteins were classi- fied according to their biological function using the Human Protein Reference Database (HPRD-http://www.hprd.org). The altered observed proteins were compared with those previously reported in proteomic studies of other brain regions of post-mortem samples in schizophrenia based on the gene symbol. PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 5 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia Immunoblotting 50 μg of total protein lysate from each sample were resolved by SDS-PAGE electrophoresis and transferred to a nitrocellulose membrane. Membranes were cut at different molecular weights and immunoblotted with rabbit polyclonal antibody against VPP1 (1:500; ab103680, Abcam) and calmodulin (1:1000; 4830, Cell Signalling Technologies), and monoclonal anti- bodies against parvalbumin (1:1000; MAB1572, Millipore-Chemicon) and glyceraldehyde- 3-phosphate dehydrogenase (GAPDH) (1:500000; MAB374, Millipore-Chemicon) to optimize the limited available brain sample. All proteins were detected by a unique band at the predicted molecular weight with the exception of VPP1 in Cohort II, where an extra lower weight molec- ular band was detected; however, it was not analysed. An extra re-incubation was needed for the parvalbumin cohort I immunoblot. Densitometric quantification of candidate proteins was performed using Quantity One software (BioRad). Values were normalized to GAPDH and a control reference sample. At least two independent immunoblot analyses were per- formed per sample. Statistical analysis For validation analysis using immunoblot we used the following procedures. Normality of the variables was assessed using the Kolmogorov-Smirnov test. Demographic and tissue-related features of the samples were compared between schizophrenia and control conditions using the Fisher exact test for qualitative variables and the unpaired Student’s t-test or the Mann- Whitney U test for quantitative parametric or non-parametric variables, respectively. The dif- ferences in protein levels between pools were analysed using the one-tailed unpaired Student’s t test, since the direction of change was expected to be the same as in the proteomics assay. The differences in protein levels between groups in the individual sample analysis were evaluated using the Kruskal-Wallis one-way analysis of variance by ranks, and the Mann-Whitney U test was used to compare differences between two groups. The Grubbs test and Pierce test were used to detect outliers for parametric or non-parametric variables, respectively; the number of outliers detected for each analysis is indicated in the figure legend. Spearman or Pearson corre- lation analyses were carried out to detect association of our molecular measures with other clinical, demographic and tissue-related variables (age, post-mortem delay, pH and toxicology, and daily antipsychotic dose, age of onset and duration of the illness in the SZ group of Cohort II). Statistical analysis was performed with GraphPad Prism version 5.00, with significance level set to 0.05. Results Proteomic analysis of post-mortem cerebellum from schizophrenia and control subjects To identify proteins significantly altered in schizophrenia in the cerebellum, we analysed the proteomes of pooled protein lysates from four subjects with schizophrenia and four control subjects matched for gender, age and post-mortem delay (S1A Fig and S1 Table). We quanti- fied 1412 proteins with a Protein Prophet probability score of more than 90% (S1 Dataset). The distribution of protein heavy-to-light (H/L) ratios revealed that some proteins of the cerebellar proteome were altered in schizophrenia (S1B and S1C Fig). We identified 99 (7%) significantly altered proteins at a false discovery rate of 10% and a protein sequence coverage greater than 5% (S2 Dataset). Moreover, 31 of the 99 significantly regulated proteins had been previously reported to be altered in proteomic analyses of other brain areas in schizophrenia (Table 2). PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 6 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia Table 2. List of altered proteins previously reported in other brain regions in proteomic studies in schizophrenia. Brain region Gene Symbol Referencea DLPFC GNB1, NDUFA12 (Behan et al., 2009) GNB1, MAG (Chan et al., 2011) VIM (English et al., 2009) GNB1, TF (Martins-de-Souza et al., 2009) ATP6V0D1, SORBS1, VIM (Martins-De-Souza et al., 2009) PNP (Novikova et al., 2006) TF (Pennington et al., 2008) BCL2L13, PHB2, SORBS1, SV2B, TMEM30A (Pinacho et al., 2016) TF (Prabakaran et al., 2004) PVALB (Smalla et al., 2008) OFC DMTN, PGAM5, VDAC2 (Velásquez et al., 2017) ACC GNB1, TF (Clark et al., 2006) AP2B1, ATP6V0A1, BAIAP2, CALM2, CCT6A, HSD17B4, NAPA, (Föcking et al., 2015) SRPRB, VDAC2 GNB1 (Martins-de-Souza et al., 2010) CC GNB1, HAPLN2, VIM (Saia-Cereda et al., 2015) BAIAP2 (Saia-Cereda et al., 2016) CALM1, CDC42 (Saia-Cereda et al., 2017) Thalamus VIM (Martins-de-Souza et al., 2010) Hippocampus PVALB (Föcking et al., 2011) ACOT7, CCT6A (Schubert et al., 2015) VIM (Nesvaderani et al., 2009) Temporal HAPLN2 (Martins-de-Souza et al., lobe 2009) CADM1, CALM1, CDC42, H3F3A, PPP3R1, VIM (Saia-Cereda et al., 2017) PHB2, VIM (MacDonald et al., 2015) DLPFC, dorsolateral prefrontal cortex; OFC, orbitofrontal cortex; ACC, anterior cingulate cortex; CC, corpus callosum. a The complete information of the references is detailed in the S1 File. https://doi.org/10.1371/journal.pone.0230400.t002 We classified the altered proteins according to their biological function and compared them to the proteins that showed no change between the two groups. Similar biological functions were found in both sets. The most prevalent function was cell communication and signalling pathways, and the transport function was enriched 10% in the non-regulated proteome and 13% in the regulated proteome (S1D Fig). We generated a list of 11 protein candidates from these functions according to two criteria: (i) the protein had been quanti- fied with more than 4 peptides; and (ii) the protein showed a greater than 2-fold increase or decrease (Table 3). Validation of protein changes in cerebellum schizophrenia samples. From this list of 11 candidates, we selected three proteins for further validation by immunoblot. We selected cal- modulin 2 (CALM2), which was identified with the highest number of peptides, and two pro- teins with the most prominent changes in each Gene Ontology function: vacuolar proton translocating ATPase 116 kDa subunit a (VPP1) from transport, and parvalbumin alpha PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 7 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia Table 3. List of proteins filtered from enriched and most representative functions in the post-mortem cerebellum in schizophrenia. Function Acc. Number Gene Symbol Protein Description (Protein Symbol) Quantified Ratio H/ log2Ratio H/ Peptides L L Transport Q93050 ATP6V0A1 V-type proton ATPase 116 kDa subunit a 5 0.09 -3.46 isoform 1 (VPP1) A0A0A0MR02 VDAC2 Voltage-dependent anion-selective channel 16 0.13 -2.97 protein 2 (VDAC2) Q86UR5 RIMS1 Regulating synaptic membrane exocytosis 11 0.46 -1.12 protein 1 (RIMS1) M0R0Y2 NAPA Alpha-soluble NSF attachment protein (SNAA) 6 3.90 1.97 Q01469 FABP5 Fatty acid-binding protein, epidermal (FABP5) 6 5.20 2.38 Cell Communication / Signal E7EMB3 CALM2 Calmodulin-2 (CALM2) 18 2.56 1.36 transduction P37235 HPCAL1 Hippocalcin-like protein 1 (HPCL1) 9 2.64 1.40 O00533 CHL1 Neural cell adhesion molecule L1-like protein 17 3.01 1.59 (NCHL1) Q8WUD1 RAB2B Ras-related protein Rab-2B (RAB2B) 6 3.17 1.66 Q96FQ6 S100A16 Protein S100-A16 (S10AG) 7 3.53 1.82 P20472 PVALB Parvalbumin alpha (PRVA) 8 3.65 1.87 Access numbers from Uniprot database Uniprot; H/L ratio between heavy (schizophrenia) and light (control) peptide areas; candidates selected for validation are shown in bold. https://doi.org/10.1371/journal.pone.0230400.t003 (PRVA) from cell communication and signal transduction (Table 3). The calmodulin antibody targets the calmodulin 1, calmodulin 2 and calmodulin 3 protein isoforms, all of which have similar molecular weight and protein sequence. We confirmed by immunoblot the changes detected by proteomics for the three candidates (suicide schizophrenia, n = 4; control, n = 4; S1 Table and S2 Fig). The fold changes of the three candidates were similar to the fold changes detected in the pooled analysis, being decreased for VPP1 and increased for calmodulin and parvalbumin (S2 Fig and Table 3). We further characterized VPP1, calmodulin and parvalbumin by immunoblot in a cohort consisting of 7 schizophrenia subjects, 7 matched control subjects and 6 matched non-schizo- phrenia suicide subjects (Table 1, Cohort I). No differences in age, post-mortem delay and pH were found between comparison groups (Table 1, Cohort I). All candidate levels were referred to GAPDH levels, which were not significantly different between the groups (Fig 1A, 1B and 1C). We found that VPP1 protein levels were again significantly decreased [t = 2.809, df = 12, p = 0.0079; Mean ± SEM: control (C) = 1.000 ± 0.143, SZ = 0.526 ± .0.089] and calmodulin and parvalbumin levels were also significantly increased in the schizophrenia group (calmodu- lin [t = 3.724, df = 12, p = 0.0029; Mean ± SEM: C = 1.000 ± 0.083, SZ = 1.918 ± 0.232], parval- bumin [t = 4.964, df = 12, p = 0.0003; Mean ± SEM: C = 1.000 ± 0.0.086, SZ = 1.524 ± 0.060] (Fig 1A, 1B and 1C). In addition, we found significant differences between the schizophrenia group and the non-schizophrenia suicide group in the fold changes for calmodulin and parval- bumin, which in the suicide group showed similar levels to the control group [calmodulin [t = 2.708, df = 11, p = 0.0204; Mean ± SEM: SZ = 1.918 ± 0.232, suicide (SC) = 1.141 ± 0.149], parvalbumin [t = 6.129, df = 11, p < 0.0001; Mean ± SEM: SZ = 1.524 ± 0.060, SC = 0.849 ± 0.096] (Fig 1A, 1B and 1C). However, the protein levels of VPP1 were not significantly differ- ent in the non-schizophrenia suicide group compared to the schizophrenia group [t = 0.3354, df = 11, p = 0.7436; Mean ± SEM: SZ = 0.526 ± 0.089, SC = 0.492 ± 0.032] (Fig 1A, 1B and 1C). Furthermore, we analysed the influence of other demographic, clinical, and tissue-related vari- ables in the differences found in the schizophrenia group compared to control subjects. None of our molecular measures showed any association with other variables of the study with the PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 8 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia Fig 1. Validation analysis of hit candidate proteins by immunoblot in Cohort I. Protein extracts from samples of the post-mortem cerebellum of non-psychiatric control (C, n = 7), schizophrenia (SZ, n = 7) and non-schizophrenia suicide (n = 6) subjects (Table 1, Cohort I) were analysed by immunoblot for VPP1, PRVA, calmodulin (CaM) and GAPDH and quantified by densitometry. (A) Protein levels for each were normalized to GAPDH values and to the mean of the control samples. Each box plot represents the median, interquartile range and range of each group from at least two independent determinations. Statistical analysis was performed using the Kruskal-Wallis test for differences between groups (VPP1: p = 0.0095; PRVA: p = 0.0011; CaM: p = 0.0270) and the t test for comparison between the indicated groups. (n.s.-not PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 9 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia significant, � p<0.05, �� p<0.01, ��� p<0.001, ���� p<0.0001). (B) Representative Western blot images for VPP1, PRVA and GAPDH in 3 non-psychiatric control individuals, 3 schizophrenia subjects and 3 non-schizophrenia suicide controls. (C) Representative Western blot images for CaM and GAPDH in 3 non-psychiatric control individuals, 3 schizophrenia subjects and 3 non-schizophrenia suicide control. See S1 Raw images for complete western blot images. https://doi.org/10.1371/journal.pone.0230400.g001 exception of parvalbumin, which showed an increase in the antipsychotic treated group compared to drug-free subjects (Table 4, Cohort I; Dunn’s Test p<0.05), indicating that the increase observed for parvalbumin in schizophrenia could be due to the antipsychotic treat- ments in these subjects. We further characterized VPP1, calmodulin and parvalbumin protein levels by immuno- blot in a larger, independent cohort of 13 non-suicide chronic schizophrenia subjects and 14 matched control individuals (Table 1, Cohort II). All candidate levels were referred to GAPDH levels, which were not significantly different between the groups (Fig 2A, 2B and 2C). We found that calmodulin and parvalbumin protein levels were significantly altered in the non-suicide schizophrenia group (calmodulin [U = 45.00, p = 0.0136; Mean ± SEM: C = 1.000 ± 0.3570, SZ = 0.6586 ± 0.2235], parvalbumin [t = 2.337, df = 25, p = 0.0139; Mean ± SEM: C = 1.000 ± 0.110, SZ = 1.380 ± 0.163]) (Fig 2A, 2B and 2C). VPP1 protein levels did not show a significant decrease in the schizophrenia group compared to the control group [t = 0.1384, df = 24, p = 0.4455; Median ± SEM: C = 0.844 ± 0.148, SZ = 0.884 ± 0.249] (Fig 2A, 2B and 2C). Association analysis of other variables in this cohort of chronic schizophrenia Table 4. Association analysis of other variables in Cohorts I and II. Cohort I Age PMD pH Toxicology r r r K C-SZ cohort (n = 14) VPP1 0.077 -0.007 -0.226 5.655 CaMa, b -0.064 -0.381 0.239 1.257 PRVA 0.112 0.006 0.071 6.600,c Cohort II Age PMD pH Arterial Hypertension r r r T; df SZ-C cohort II (n = 27) CaMa 0.221 -0.171 -0.048 1.13; 25 PRVA 0.032 0.208 0.169 0.55; 25 Daily AP dosed Age of onset Duration of illness Body mass index Medication side-effectse r r r’ r r’ SZ cohort II (n = 13) CaM -0.116 -0.099 0.124 -0.170 0.062 PRVA 0.162 -0.170 -0.028 -0.231 0.189 r, Pearson’s r for parametric variables; r’, Spearman’s correlation for non-parametric variables; PMD, post-mortem delay; C, control; SZ, schizophrenia; AP, antipsychotic. a r of Spearman are shown for this variable. b An outlier was detected for calmodulin (CaM) and therefore excluded from the SZ-C analysis (CaM: C, n = 13, SZ, n = 13). c p<0.05. d Last chlorpromazine equivalent dose was calculated based on the electronic records of drug prescriptions of the patients. e Tardive Dyskinesia assessed by Abnormal Involuntary Movement Scale; K, Kruskal-Wallis; T, T-statistic; df, degrees of freedom; N/A, not applicable. Significant associations are indicated in bold. https://doi.org/10.1371/journal.pone.0230400.t004 PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 10 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia Fig 2. Validation analysis of hit candidate proteins by immunoblot in Cohort II. Protein extracts from samples of the post-mortem cerebellum of non-suicide chronic schizophrenia (SZ, n = 13) and control (C, n = 14) subjects (Table 1, Cohort II) were analysed by immunoblot for the same proteins as in Fig 1A and quantified by densitometry. (A) Protein levels for each protein were normalized to GAPDH values and to the mean of the control samples. Each box plot represents the median, interquartile range, and range of each group from at least two independent determinations. An outlier was detected for VPP1 protein values in the control group. Statistical analysis for comparison between case and control groups was performed using the t test for VPP1 and PRVA and the Mann- Whitney U test for CaM and GAPDH (n.s.-not significant, � p<0.05, ��� p<0.001). (B) Representative Western blot images for VPP1, PRVA, CaM and GAPDH in 6 non-psychiatric control individuals and 6 non-suicide schizophrenia subjects from set 2. See S2 Raw images for complete western blot images. https://doi.org/10.1371/journal.pone.0230400.g002 PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 11 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia (age, post-mortem delay, pH, arterial hypertension, antipsychotic dose, age of onset, duration of the illness, body mass index, and medication side-effects) did not show any associations between calmodulin and parvalbumin protein levels and other variables (Table 4, Cohort II). Discussion Proteomic screening of cerebellum samples from schizophrenia patients has helped to identify candidates with a putative relevant role in this disorder. The list of altered proteins in this region in suicide schizophrenia subjects provides the first step to analyse in depth some of the most robust candidates. A systematic validation of three candidates in two independent cohorts, one containing suicide individuals and another containing schizophrenia subjects that died of natural causes, revealed that only the two EF-hand calcium-binding proteins were altered in schizophrenia but not in non-schizophrenia suicide subjects, while the third candi- date, a subunit of the proton pump ATPase, was linked to suicide behaviours and was not dys- regulated in non-suicidal schizophrenia subjects. Thus, our findings propose that the EF-hand calcium-binding proteins parvalbumin and calmodulin could be involved in the pathophysiol- ogy of schizophrenia in the cerebellum by altering calcium-dependent signalling pathways involved in synaptic function. Moreover, this work provides evidence for the role of VPP1 in suicide behaviours with a possible impact on the synaptic vesicle cycle. Alterations in both the calcium-binding proteins in schizophrenia and in the proton pump ATPase in suicide behav- iours could lead to a disruption in the cerebellar synaptic functioning of the neuronal circuits through different mechanisms as discussed below. Vacuolar proton ATPase pump Here we describe a decrease in protein levels of the subunit of the vacuolar-type proton pump ATPase (VPP1) in suicide subjects with schizophrenia or other psychiatric disorders, but not in non-suicidal elderly schizophrenia patients, suggesting that this protein may be involved in suicide behaviours rather than in schizophrenia. Another report has also found a decrease in VPP1 in younger schizophrenia subjects in the anterior cingulated cortex using the Stanley Medical Research Institute’s (SMRI) Array Collection (http://www.stanleyresearch.org), which included 4 suicides out of 15 subjects [30]. In line with our results, VPP1 has been reported in a list of genes differently expressed in the frontal cortex in subjects with major depression that committed suicide [31]. VPP1 is part of the large multi-subunit complex of the vacuolar pro- ton pump ATPase. This complex provides a proton motive force and is involved in many cel- lular and physiological functions (reviewed in [32]). V+H-ATPase pumps are composed of two functional components, the cytoplasmic V1 component with the catalytic activity for ATP hydrolysis and the V0 component, which forms a membrane embedded component that is required for proton translocation across the membrane, which occurs through the a subunit (reviewed in [32]). The VPP1 subunit is highly expressed in neuronal cells. This subunit is a major component of synaptic vesicles and provides the pH gradient and membrane potential required for neurotransmitter accumulation in the initial phase of the synaptic vesicle cycle [33,34]. Later studies provided evidence that subunit a of the V0 domain is also involved in synaptic vesicle release by mediating the membrane fusion events downstream of the t- SNARE docking of vesicles in a calcium/calmodulin-dependent manner [35,36]. Indeed, the VPP1 orthologue in fly neurons directly interacts with calmodulin, another protein that we found altered in schizophrenia (see above), and this interaction is required for recruiting cal- modulin to synapses and for the viability function of VPP1 [35,37]. In addition, neurons lack- ing VPP1 in Drosophila melanogaster and Caenorhabditis elegans accumulate vesicles in the synaptic terminals, supporting the role of this protein in neurotransmitter release into the PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 12 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia synaptic cleft [35,38,39]. Thus, the reduction of VPP1 that we found in the cerebellum in sui- cide subjects could be impairing normal neurotransmitter accumulation in synaptic vesicles and their subsequent release into the synaptic terminals. A limited synaptic response of the cerebellar neurons could contribute to the altered functioning of cerebellar circuits in suicide behaviours. In agreement with this idea of altered synaptic activity in suicide are the findings of multiple studies. The regulation of the expression of genes involved in neurotransmission and synaptic function has been linked to suicide behaviours in different psychiatric disorders, including schizophrenia and major depression [31,40–43]. For example, increased expression of the 5-hydroxytryptamine (5-HT) receptor type 2A has been reported in the amygdala of suicide subjects together with increased 5-HT utilisation [44–46] as well as in other brain areas such as the hippocampus and the prefrontal cortex [47,48], suggesting a widespread dysregulation of the serotonin system in suicide subjects, which has led to numerous studies of serotonin func- tion in the context of suicidality [49,50]. Thus, changes in intracellular proteins involved in neurotransmitter trafficking are likely to have a widespread effect on this neurotransmitter system. Interestingly, the cerebellum has been described as one of the neural routes altered in suicide behaviours related to monoaminergic signal transduction [51]. There is evidence of altered subcellular localisation of 5-HT2A receptors in Purkinje cells as well as an increase in white matter in the cerebellum in schizophrenia [52], suggesting a possible compensatory mechanism for compromised 5-HT synthesis or release in this region. Some reports have also linked reduced cerebellar activity and reduced grey matter with the ideation of suicide or sui- cide attempt, respectively [53,54]. Thus, the reduction of VPP1 observed in our study in sui- cide subjects in the cerebellum could alter the refilling of monoaminergic neurotransmitters (5-HT mainly) into synaptic vesicles and their release in cerebellar circuits. Moreover, it might mediate the decreased activity reported in this area in suicide behaviours. EF-hand calcium-binding proteins Calcium homeostasis has been suggested to be disrupted in schizophrenia [55,56]. In our study, we found altered levels of two proteins that are sensitive to changes in the intracellular concentrations of calcium and that play a key role in signalling transduction and synaptic functioning. Calmodulin. Calmodulin (CaM) belongs to the large family of EF-hand calcium-binding proteins. Here, we report an increase in calmodulin levels in the cerebellum in schizophrenia in the chronic SZ group with a mean age of 44 years, while calmodulin was found to be down- regulated in a late elderly chronic schizophrenia cohort (mean age of 74 years). A previous study in an early elderly chronic schizophrenia cohort (mean age of 68) also reported a down- regulation of calmodulin protein levels in different brain areas [57], while another study of chronic schizophrenia (mean age 67) reported an upregulation of calmodulin protein levels in nuclear-enriched cell samples from the corpus callosum and anterior temporal lobe [58]. In line with our study, these reports suggest a different regulation of calmodulin that may depend on age, brain region, or the cellular or subcellular locations under study. Calmodulin is the major calcium-binding protein present in the brain and acts as a calcium sensor, detecting and responding to biologically relevant changes in intracellular concentrations of calcium [59–61]. Calcium regulates calmodulin by changing its subcellular localization, promoting interaction with many proteins or by inducing conformational changes that allow the interaction and acti- vation of specific targets and the subsequent triggering of a signalling cascade [59]. Calmodu- lin-dependent kinase II (CaMKII) has been proposed as a susceptibility gene in schizophrenia [62]. CaMKII is an important calmodulin effector in neurons that has been implicated in PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 13 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia activity-dependent functions such as gene transcription, signalling and synaptic and dendritic development, maturation and function as well as in cognition [60,63]. The list of calmodulin effectors is extensive and includes plasma membrane calcium pumps, various ion channels, protein kinases and receptors, among others [59,61,64]. Calcium homeostasis is thus tightly regulated and features downstream of some of the mechanisms suggested to be altered in suicide subjects, such as an increase in 5-HT2A signalling, whose downstream intracellular cascades converge in the release of intracellular calcium. Thus, the calmodulin-dependent functions that could be dysregulated in schizophrenia in the cerebellum are wide, leading to an important impact on cerebellar circuit functioning and formation in accordance with the connectivity deficits and the neurodevelopmental hypothesis for schizophrenia [1,65]. In our study, we also found in suicide schizophrenia subjects a decrease in one interactor of calmodu- lin in flies, the orthologue of VPP1, which is required for recruiting calmodulin to synapses [37,66]. Indeed, the Ca2+-CaM regulation of V100 (VPP1 orthologue in flies) has been pro- posed as a positive regulator of the assembly of the SNARE complex on distinct vesicles and subsequent neurotransmitter release [37]. If this interaction is present in cerebellar neurons, this evidence could either suggest that the upregulation of calmodulin in the cerebellum in sui- cide schizophrenia subjects could be a mechanism to compensate for the possible lower abun- dance of calmodulin at the synapses due to a reduction in VPP1 levels or that it could be a compensatory mechanism for the lower formation/release of synaptic vesicles [37,66]. Further investigations will be needed to confirm this possibility. However, in chronic SZ patients that died from natural causes, calmodulin was decreased, suggesting a possible constitutive down- regulation of calmodulin in chronic schizophrenia. In this context, VPP1 levels were not altered, raising the possibility that the recruitment of calmodulin to the synapse could be cor- rect and the reduction of calmodulin could be impacting on other calmodulin-dependent functions in the cerebellum. Parvalbumin. Here, we report an increase in parvalbumin in the cerebellum in schizo- phrenia independently of the mechanism of death. Parvalbumin, like calmodulin, also belongs to the large family of EF-hand calcium-binding proteins. However, parvalbumin is a calcium buffer protein, which are proteins essential for modulating calcium homeostasis in neurons and are implicated in the subtle regulation and timing of calcium signals pre- and post-synap- tically [67–69]. Parvalbumin is a spatial and temporal regulator of calcium transients that mod- ulates calcium pools and that is critical for synaptic plasticity, such as short-term facilitation [67,68]. Furthermore, it has been proposed that it could also regulate calcium signalling as a slow-onset calcium sensor in addition to being regulated by calcium [70]. Parvalbumin is expressed in subpopulations of GABAergic interneurons in the brain, which are considered more metabolically and electrically active than the neighbouring neurons and play an impor- tant role in the pathophysiology of schizophrenia [14,71,72]. Indeed, a decrease in the expres- sion of parvalbumin in patients with SZ in different brain regions including the hippocampus and prefrontal cortex has been a consistent finding in human post-mortem studies [56,73] and in animal models of schizophrenia [74–77]. However, little is known about parvalbumin pro- tein levels in the cerebellum of SZ patients. In the cerebellum, parvalbumin localizes to the axon, soma and dendrites of Purkinje, stellate, basket and a small proportion of Golgi cells [70]. Studies in parvalbumin-deficient mice suggest that this protein is required for normal locomotor activity in order to maintain the normal spontaneous arrhythmic and asynchro- nous firing pattern of Purkinje cells [78,79]. These mice also showed behavioural deficits linked to schizophrenia and autism, such as deficits in sensorimotor gating and novelty seek- ing and reduced social interaction and communication [80,81]. The increase in parvalbumin we observed in the cerebellum in schizophrenia could produce an imbalance in the regulation of intracellular calcium concentration and/or altering calcium-dependent signalling during PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 14 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia synapse activity and this could have an impact on the behavioural changes observed in schizo- phrenia. The increase in parvalbumin could also reflect a change in the number of certain sub- populations of parvalbumin-positive neurons in the cerebellum in schizophrenia. However, a reduction in Purkinje neurons in the cerebellum has been reported in subjects with schizo- phrenia [72], suggesting that this possibility may be more likely due to the contribution from other types of parvalbumin-expressing neurons in the cerebellum. Further studies will be needed to elucidate these possibilities. Functional neuroimaging studies show a predominantly hypoactivation of the cerebellum in schizophrenia [82,83]. Thus, based on our results, the altered EF-hand calcium-binding pro- teins found in the cerebellum could have an impact on synaptic transmission and underlie the reduced cerebellar activity observed in people with schizophrenia. Further studies are needed to investigate this hypothesis. Limitations We acknowledge several limitations of our study. First, we used pooled samples in the proteo- mic screening. Although this type of design is a useful approach for detecting commonly altered pathways [84–88], it precludes the discovery of individual-specific changes or control- ling for inter-individual variations. In our immunoblot analysis, the results obtained with pooled samples were recapitulated in individual samples from a similar cohort, suggesting that molecular changes in the cerebellum could be conserved across individuals. Second, the sam- ple size of Cohort I is limited. Further analysis in a larger, independent cohort of samples including suicidal groups will be needed to explore how stable our findings are in other patients. Third, antipsychotic treatments could influence the results. To control for this vari- able, we have used the blood toxicology data in Cohort I and chlorpromazine equivalent daily dose data in Cohort II, showing that the increased expression of parvalbumin could be due to the antipsychotic treatments in Cohort I but not in Cohort II. Further pharmacological studies in cellular and animal models, as well as in drug-naive patients, will help to clarify the effect of antipsychotics on parvalbumin. Moreover, no information about the duration of the treat- ments was available for the schizophrenia groups. Thus, the effect of long-term treatments on our molecular variables could not be assessed. Fourth, the schizophrenia subjects used in our proteomic screening committed suicide, which could be influencing the findings observed. We have controlled for this possibility in the validation of candidates by including a group with subjects who committed suicide but with varying psychiatric diagnoses instead of schizo- phrenia and an independent cohort with non-suicide schizophrenia subjects that died from natural causes. These analyses have allowed us to detect that VPP1 alteration could be a com- mon feature of suicide and suggest that this candidate may only be altered in suicide subjects rather than in schizophrenia. It may also be possible that alterations of these proteins could occur in other psychiatric disorders. Further studies of these candidates in anxiety and adjust- ment disorders will be helpful to explore this possibility. Last, the majority of the subjects included in both cohorts were male. Further studies in a cohort with equal representation of both genders would be of interest. Conclusions In summary, our findings provide evidence for an upregulation of calcium-binding proteins in the cerebellum in schizophrenia together with a calmodulin downregulation in chronic schizo- phrenia in elderly subjects, suggesting an altered modulation of calcium signalling and calcium transients in synaptic responses in this region in schizophrenia with an impact on the normal functioning of cerebellar circuits. In addition, our study provides evidence for the alteration of PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 15 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia VPP1 in the cerebellum linked to suicide behaviours, suggesting an involvement of defective synaptic vesicle cycle and the release of neurotransmitters in suicide behaviours. Supporting information S1 Fig. Experimental strategy for large-scale quantitative proteomic analysis and identifi- cation of differentially expressed proteins in cerebellum in schizophrenia. (A) Protein lysates from the post-mortem cerebellum of control (n = 4) and suicide schizophrenia (SZ, n = 4) subjects (Table 1) were processed as described in the experimental procedures section. In the analysis in pools, samples from the same group were pooled. In the analysis of individual samples from schizophrenia, a pool of controls was used to compare each individual sample of schizophrenia. Subsequently, protein database searches, peptide quantification and data analy- sis were performed as described in the experimental procedures section. A panel of 3 candi- dates from significantly regulated proteins was selected for further validation by immunoblot: first, in a pilot cohort which includes a group of non-schizophrenia suicide subjects (Table 1; Cohort I: SZ (n = 7), non-SZ suicide (n = 6), control (n = 7)) and then in a larger cohort of non-suicide chronic schizophrenia subjects (Table 1, Cohort II: non-suicide SZ (n = 13), control (n = 14)). (B) Distribution of the number of peptides quantified per protein from the data set of 2289 quantified proteins. (C) Normalized distribution of z-scores for confidently quantified proteins (>2 peptide sequences) (n = 1148). (D) Gene ontology classification of biological functions for non-significantly and significantly altered proteins with low variation in the cerebellum in SZ compared to the control. Transport (GO:0006810); Cell communica- tion (GO:0007154); Signal transduction (GO:0007165); Metabolism (GO:0008152); Energy pathways (GO:0006091); Regulation of nucleobase, nucleoside, nucleotide and nucleic acid metabolism (GO:0019219); Cell growth and/or maintenance (GO:0008151); Protein metabo- lism (GO:0019538); Biological process unknown (GO:0000004). (TIF) S2 Fig. Validation of hit candidate proteins by immunoblot in pools. Pooled protein extracts from samples of the post-mortem cerebellum of control (C, n = 4) and schizophrenia (SZ, n = 4) subjects from the UPV/EHU brain collection (S1 Table, a subgroup from Cohort I, Table 1) used in the proteomic screening were analysed by immunoblotting for VPP1, PRVA, calmodulin (CaM) and GAPDH. Protein levels for each hit were quantified by densitometry and normalized to GAPDH values and to the reference control sample. Images show represen- tative immunoblots of a pool of control (left band, C) and a pool of schizophrenia (right band, SZ) subjects. Analysis was performed in duplicate. Bars represent mean ± standard deviation of the analysis of duplicates from two independent dissections, with the exception of PVALB, whose data are from a duplicate analysis of one dissection. Statistical analysis was performed using the t test (n.s.-not significant, �� p<0.01, ��� p<0.001). (TIF) S1 Raw images. Validation analysis of hit candidate proteins by immunoblot in Cohort I. Protein extracts from samples of the post-mortem cerebellum of non-psychiatric control (C, n = 7), schizophrenia (SZ, n = 7) and non-schizophrenia suicide (n = 6) subjects (Table 1, Cohort I) were analysed by immunoblot for VPP1, PRVA, calmodulin (CaM) and GAPDH and quantified by densitometry. Images show uncropped images of the area of the membrane incubated with anti-VPP1, anti- parvalbumin (PRVA) (A), anti-CaM (B) and anti-GAPDH (A and B) of immunoreactivities of Fig 1. The samples shown in Fig 1 are delimited by a dashed line on the complete Western blot membranes. Arrows indicate the analysed band. X, sample PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 16 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia not included in Fig 1. � , Non-analysed immunoreactivity. (TIF) S2 Raw images. Validation analysis of hit candidate proteins by immunoblot in Cohort II. Protein extracts from samples of the post-mortem cerebellum of non-psychiatric control (C, n = 14) and schizophrenia (SZ, n = 13) subjects (Table 1, Cohort II) were analysed by immu- noblot for VPP1, PRVA, CaM and GAPDH and quantified by densitometry. Images show uncropped images of the area of the membrane incubated with anti-VPP1, anti-PRVA, anti- CaM and anti-GAPDH of immunoreactivities of Fig 2. The samples shown in Fig 2 are delim- ited by a dashed line on the complete Western blot membrane. Arrows indicate the analysed band. X, sample not included in the Fig 2. � , Non-analysed immunoreactivity. (TIF) S1 Table. Demographic, clinical and tissue-related features of cases used for quantitative proteomic analysis. Mean ± standard deviation or relative frequency are shown for each vari- able; PMD, post-mortem delay; SZ, schizophrenia; C, healthy control group; AP, antipsychot- ics; N/A, not applicable. 1Paranoid schizophrenia (n = 7). 2Mann-Whitney U is shown for non-parametric variables. (DOCX) S1 Dataset. List of reliably quantified proteins in the cerebellum in schizophrenia. (Proba- bility >90%). (XLSX) S2 Dataset. Proteins significantly regulated in the cerebellum in schizophrenia, classified according to their biological function (FDR<0.1, coverage >5%). (XLSX) S1 File. Supplementary material and methods. (DOCX) Acknowledgments The authors thank the donors and their families for the donation of their brains; the technical support of Nuria Villalmanzo; the staff members of the Basque Institute of Legal Medicine for their help; and the English language correction of Dr Rose. Author Contributions Conceptualization: Belén Ramos. Formal analysis: Francisco Vidal-Domènech, Gemma Riquelme, Raquel Pinacho, Ricard Rodriguez-Mias, América Vera, Judit Villén, Belén Ramos. Funding acquisition: Belén Ramos. Investigation: Francisco Vidal-Domènech, Gemma Riquelme, Raquel Pinacho, Ricard Rodri- guez-Mias, Judit Villén, Belén Ramos. Methodology: Ricard Rodriguez-Mias, Judit Villén, Belén Ramos. Resources: Alfonso Monje, Isidre Ferrer, Luis F. Callado, J. Javier Meana. Supervision: Belén Ramos. PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 17 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia Writing – original draft: Belén Ramos. Writing – review & editing: Francisco Vidal-Domènech, Raquel Pinacho. References 1. Andreasen NC, Nopoulos P, O’Leary DS, Miller DD, Wassink T, Flaum M. Defining the phenotype of schizophrenia: cognitive dysmetria and its neural mechanisms. Biol Psychiatry. 1999; 46: 908–20. https://doi.org/10.1016/s0006-3223(99)00152-3 PMID: 10509174 2. Parker KL, Kim YC, Kelley RM, Nessler AJ, Chen K-H, Muller-Ewald VA, et al. Delta-frequency stimula- tion of cerebellar projections can compensate for schizophrenia-related medial frontal dysfunction. Mol Psychiatry. 2017; 22: 647–655. https://doi.org/10.1038/mp.2017.50 PMID: 28348382 3. Brady RO, Gonsalvez I, Lee I, Öngür D, Seidman LJ, Schmahmann JD, et al. Cerebellar-prefrontal net- work connectivity and negative symptoms in schizophrenia. Am J Psychiatry. 2019. https://doi.org/10. 1176/appi.ajp.2018.18040429 PMID: 30696271 4. Adamaszek M, D’Agata F, Ferrucci R, Habas C, Keulen S, Kirkby KC, et al. Consensus paper: cerebel- lum and emotion. The Cerebellum. 2017; 16: 552–576. https://doi.org/10.1007/s12311-016-0815-8 PMID: 27485952 5. Andreasen NC, Pierson R. The Role of the cerebellum in schizophrenia. Biol Psychiatry. 2008; 64: 81– 88. https://doi.org/10.1016/j.biopsych.2008.01.003 PMID: 18395701 6. Sokolov AA, Miall RC, Ivry RB. The cerebellum: Adaptive prediction for movement and cognition. Trends Cogn Sci. 2017; 21: 313–332. https://doi.org/10.1016/j.tics.2017.02.005 PMID: 28385461 7. Mothersill O, Knee-Zaska C, Donohoe G. Emotion and theory of mind in schizophrenia—investigating the role of the cerebellum. The Cerebellum. 2016; 15: 357–368. https://doi.org/10.1007/s12311-015- 0696-2 PMID: 26155761 8. Ben-Yehudah G, Guediche S, Fiez JA. Cerebellar contributions to verbal working memory: beyond cog- nitive theory. The Cerebellum. 2007; 61: 193–201. https://doi.org/10.1080/14734220701286195 PMID: 17786815 9. Bellebaum C, Daum I. Cerebellar involvement in executive control. The Cerebellum. 2007; 6: 184–192. https://doi.org/10.1080/14734220601169707 PMID: 17786814 10. Ackermann H, Mathiak K, Riecker A. The contribution of the cerebellum to speech production and speech perception: Clinical and functional imaging data. The Cerebellum. 2007; 6: 202–213. https://doi. org/10.1080/14734220701266742 PMID: 17786816 11. Beckinghausen J, Sillitoe RV. Insights into cerebellar development and connectivity. Neuroscience Letters. Elsevier Ireland Ltd; 2019. pp. 2–13. https://doi.org/10.1016/j.neulet.2018.05.013 PMID: 29746896 12. Sillitoe RV, Lackey EP, Heck DH. Recent advances in understanding the mechanisms of cerebellar granule cell development and function and their contribution to behavior. F1000Research. F1000 Research Ltd; 2018. p. 1142. https://doi.org/10.12688/f1000research.15021.1 PMID: 30109024 13. Fatemi SH, Folsom TD. GABA receptor subunit distribution and FMRP–mGluR5 signaling abnormali- ties in the cerebellum of subjects with schizophrenia, mood disorders, and autism. Schizophr Res. 2015; 167: 42–56. https://doi.org/10.1016/j.schres.2014.10.010 PMID: 25432637 14. Fatemi SH, Folsom TD, Rooney RJ, Thuras PD. Expression of GABAA α2-, β1- and ε-receptors are altered significantly in the lateral cerebellum of subjects with schizophrenia, major depression and bipolar disorder. Transl Psychiatry. 2013; 3: e303–e303. https://doi.org/10.1038/tp.2013.64 PMID: 24022508 15. Burnet P, Hutchison L, Vonhesling M, Gilbert E, Brandon N, Rutter A, et al. Expression of D-serine and glycine transporters in the prefrontal cortex and cerebellum in schizophrenia. Schizophr Res. 2008; 102: 283–294. https://doi.org/10.1016/j.schres.2008.02.009 PMID: 18400471 16. Schmitt A, Koschel J, Zink M, Bauer M, Sommer C, Frank J, et al. Gene expression of NMDA receptor subunits in the cerebellum of elderly patients with schizophrenia. Eur Arch Psychiatry Clin Neurosci. 2010; 260: 101–111. https://doi.org/10.1007/s00406-009-0017-1 PMID: 19856012 17. Pinacho R, Villalmanzo N, Roca M, Iniesta R, Monje A, Haro JM, et al. Analysis of Sp transcription fac- tors in the postmortem brain of chronic schizophrenia: A pilot study of relationship to negative symp- toms. J Psychiatr Res. 2013; 47: 926–934. https://doi.org/10.1016/j.jpsychires.2013.03.004 PMID: 23540600 18. Keller S, Punzo D, Cuomo M, Affinito O, Coretti L, Sacchi S, et al. DNA methylation landscape of the genes regulating D-serine and D-aspartate metabolism in post-mortem brain from controls and subjects PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 18 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia with schizophrenia. Sci Rep. 2018; 8: 10163. https://doi.org/10.1038/s41598-018-28332-x PMID: 29976992 19. Mudge J, Miller NA, Khrebtukova I, Lindquist IE, May GD, Huntley JJ, et al. Genomic convergence anal- ysis of schizophrenia: mRNA sequencing reveals altered synaptic vesicular transport in post-mortem cerebellum. PLoS One. 2008; 3: e3625. https://doi.org/10.1371/journal.pone.0003625 PMID: 18985160 20. Kim S, Cho H, Lee D, Webster MJ. Association between SNPs and gene expression in multiple regions of the human brain. Transl Psychiatry. 2012; 2: e113. https://doi.org/10.1038/tp.2012.42 PMID: 22832957 21. Kim S, Webster MJ. The Stanley Neuropathology Consortium Integrative Database: a novel, web- Based tool for exploring neuropathological markers in psychiatric disorders and the biological processes Associated with Abnormalities of Those Markers. Neuropsychopharmacology. 2010; 35: 473–482. https://doi.org/10.1038/npp.2009.151 PMID: 19829293 22. Roca Casasús M, Escanilla Casal A, Monje Hernández A, Baño Galindo V, Planchat Teruel LM, Costa Escola J, et al. Banco de tejidos neurológicos de Sant Joan de Déu-Serveis de Salut Mental para la investigación de las enfermedades mentales. La importancia de un programa de donación en vida. Psi- quiatr Biológica. 2008; 15: 73–79. https://doi.org/10.1016/S1134-5934(08)71126-6 23. Gardner DM, Murphy AL, O’Donnell H, Centorrino F, Baldessarini RJ. International consensus study of antipsychotic dosing. Am J Psychiatry. 2010; 167: 686–693. https://doi.org/10.1176/appi.ajp.2009. 09060802 PMID: 20360319 24. Burti L, Parolin A, Zanotelli R. [Tardive Dyskinesia. AIMS (Abnormal Involuntary Movement Scale) as a Diagnostic and Research Tool]. 1981; 2829–2836. https://pubmed.ncbi.nlm.nih.gov/7312199/ 25. Pinacho R, Villalmanzo N, Lalonde J, Haro JM, Meana JJ, Gill G, et al. The transcription factor SP4 is reduced in postmortem cerebellum of bipolar disorder subjects: control by depolarization and lith- ium. Bipolar Disord. 2011; 13: 474–485. https://doi.org/10.1111/j.1399-5618.2011.00941.x PMID: 22017217 26. Khidekel N, Ficarro SB, Clark PM, Bryan MC, Swaney DL, Rexach JE, et al. Probing the dynamics of O- GlcNAc glycosylation in the brain using quantitative proteomics. Nat Chem Biol. 2007; 3: 339–348. https://doi.org/10.1038/nchembio881 PMID: 17496889 27. Vizcaı́no JA, Csordas A, Del-Toro N, Dianes JA, Griss J, Lavidas I, et al. 2016 update of the PRIDE database and its related tools. Nucleic Acids Res. 2016; 44: D447–D456. https://doi.org/10.1093/nar/ gkv1145 PMID: 26527722 28. Graumann J, Hubner NC, Kim JB, Ko K, Moser M, Kumar C, et al. Stable isotope labeling by amino acids in cell culture (SILAC) and proteome quantitation of mouse embryonic stem cells to a depth of 5,111 proteins. Mol Cell Proteomics. 2008; 7: 672–683. https://doi.org/10.1074/mcp.M700460-MCP200 PMID: 18045802 29. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to mul- tiple testing. J Roy Stat Soc Ser B. 1995. pp. 289–300. https://doi.org/10.1111/j.2517-6161.1995. tb02031.x 30. Föcking M, Lopez LM, English JA, Dicker P, Wolff A, Brindley E, et al. Proteomic and genomic evidence implicates the postsynaptic density in schizophrenia. Mol Psychiatry. 2015; 20: 424–432. https://doi. org/10.1038/mp.2014.63 PMID: 25048004 31. Zhurov V, Stead JDH, Merali Z, Palkovits M, Faludi G, Schild-Poulter C, et al. Molecular pathway recon- struction and analysis of disturbed gene expression in depressed individuals who died by suicide. Yara- gudri VK, editor. PLoS One. 2012; 7: e47581. https://doi.org/10.1371/journal.pone.0047581 PMID: 23110080 32. Maxson ME, Grinstein S. The vacuolar-type H+-ATPase at a glance—more than a proton pump. J Cell Sci. 2014; 127: 4987–4993. https://doi.org/10.1242/jcs.158550 PMID: 25453113 33. Moriyama Y, Futai M. H(+)-ATPase, a primary pump for accumulation of neurotransmitters, is a major constituent of brain synaptic vesicles. Biochem Biophys Res Commun. 1990; 173: 443–8. https://doi. org/10.1016/s0006-291x(05)81078-2 PMID: 1979489 34. Takamori S, Holt M, Stenius K, Lemke EA, Grønborg M, Riedel D, et al. Molecular Anatomy of a Traf- ficking Organelle. Cell. 2006; 127: 831–846. https://doi.org/10.1016/j.cell.2006.10.030 PMID: 17110340 35. Morel N, Poëa-Guyon S. The membrane domain of vacuolar H+ATPase: a crucial player in neurotrans- mitter exocytotic release. Cell Mol Life Sci. 2015; 72: 2561–2573. https://doi.org/10.1007/s00018-015- 1886-2 PMID: 25795337 36. Bayer MJ, Reese C, Bühler S, Peters C, Mayer A. Vacuole membrane fusion: V0 functions after trans- SNARE pairing and is coupled to the Ca2+-releasing channel. J Cell Biol. 2003; 162: 211–222. https:// doi.org/10.1083/jcb.200212004 PMID: 12876274 PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 19 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia 37. Wang D, Epstein D, Khalaf O, Srinivasan S, Williamson WR, Fayyazuddin A, et al. Ca2+–Calmodulin regulates SNARE assembly and spontaneous neurotransmitter release via v-ATPase subunit V0a1. J Cell Biol. 2014; 205: 21–31. https://doi.org/10.1083/jcb.201312109 PMID: 24733584 38. Hiesinger PR, Fayyazuddin A, Mehta SQ, Rosenmund T, Schulze KL, Zhai RG, et al. The v-ATPase V0 subunit a1 is required for a late step in synaptic vesicle exocytosis in Drosophila. Cell. 2005; 121: 607– 620. https://doi.org/10.1016/j.cell.2005.03.012 PMID: 15907473 39. Liégeois S, Benedetto A, Garnier JM, Schwab Y, Labouesse M. The V0-ATPase mediates apical secre- tion of exosomes containing Hedgehog-related proteins in Caenorhabditis elegans. J Cell Biol. 2006; 173: 949–961. https://doi.org/10.1083/jcb.200511072 PMID: 16785323 40. Roy B, Dwivedi Y. Understanding the neuroepigenetic constituents of suicide brain. Progress in molecu- lar biology and translational science. 2018. pp. 233–262. https://doi.org/10.1016/bs.pmbts.2018.01.007 PMID: 29933952 41. Serafini G, Pompili M, Innamorati M, Giordano G, Montebovi F, Sher L, et al. The role of microRNAs in synaptic plasticity, major affective disorders and suicidal behavior. Neurosci Res. 2012; 73: 179–190. https://doi.org/10.1016/j.neures.2012.04.001 PMID: 22521503 42. Bortolato M, Pivac N, Muck Seler D, Nikolac Perkovic M, Pessia M, Di Giovanni G. The role of the sero- tonergic system at the interface of aggression and suicide. Neuroscience. 2013; 236: 160–85. https:// doi.org/10.1016/j.neuroscience.2013.01.015 PMID: 23333677 43. Perenyi A, Forlano R. Suicide in schizophrenia. Neuropsychopharmacol Hung. 2005; 7: 107–17. https:// doi.org/10.1586/ern.10.82 PMID: 16493874 44. Parsey RV., Hastings RS., Oquendo MA., Huang YY, Simpson N, Arcement J, et al. Lower serotonin transporter binding potential in the human brain during major depressive episodes. Am J Psychiatry. 2006; 163: 52–58. https://doi.org/10.1176/appi.ajp.163.1.52 PMID: 16390889 45. Hrdina PD, Demeter E, Vu TB, Sótónyi P, Palkovits M. 5-HT uptake sites and 5-HT2 receptors in brain of antidepressant-free suicide victims/depressives: increase in 5-HT2 sites in cortex and amygdala. Brain Res. 1993; 614: 37–44. https://doi.org/10.1016/0006-8993(93)91015-k PMID: 8348328 46. Anisman H, Du L, Palkovits M, Faludi G, Kovacs GG, Szontagh-Kishazi P, et al. Serotonin receptor sub- type and p11 mRNA expression in stress-relevant brain regions of suicide and control subjects. J Psy- chiatry Neurosci. 2008; 33: 131–141. PMID: 18330459 47. Pandey GN, Dwivedi Y, Rizavi HS, Ren X, Pandey SC, Pesold C, et al. Higher expression of serotonin 5-HT2A receptors in the postmortem brains of teenage suicide victims. Am J Psychiatry. 2002; 159: 419–429. https://doi.org/10.1176/appi.ajp.159.3.419 PMID: 11870006 48. Turecki G, Brière R, Dewar K, Antonetti T, Lesage AD, Séguin M, et al. Prediction of level of seroto- nin 2A receptor binding by serotonin receptor 2A genetic variation in postmortem brain samples from subjects who did or did not commit suicide. Am J Psychiatry. 1999; 156: 1456–1458. PMID: 10484964 49. Arango V, Huang YY, Underwood MD, Mann JJ. Genetics of the serotonergic system in suicidal behav- ior. J Psychiatr Res. 2003; 37: 375–386. https://doi.org/10.1016/s0022-3956(03)00048-7 PMID: 12849930 50. Underwood MD, Kassir SA, Bakalian MJ, Galfalvy H, Dwork AJ, Mann JJ, et al. Serotonin receptors and suicide, major depression, alcohol use disorder and reported early life adversity. Transl Psychiatry. 2018; 8: 1–15. 51. Bozorgmehr A., Bozorgmehr A, Ghadirivasfi M, Tavakoli M, Rahmani H, Heydari F SAEI analysis of the genetic basis of suicidal behavior. PG 2018; 28: 31–37., Ghadirivasfi M, Tavakoli M, Rahmani H, Hey- dari F, Shahsavand Ananloo E. Integrated analysis of the genetic basis of suicidal behavior. Psychiatr Genet. 2018;28: 31–37. 52. Eastwood SL, Burnet PWJ, Gittins R, Baker K, Harrison PJ. Expression of serotonin 5-HT2A receptors in the human cerebellum and alterations in schizophrenia. Synapse. 2001; 42: 104–114. https://doi.org/ 10.1002/syn.1106 PMID: 11574947 53. Lee YJ, Kim S, Gwak AR, Kim SJ, Kang S-G, Na K-S, et al. Decreased regional gray matter volume in suicide attempters compared to suicide non-attempters with major depressive disorders. Compr Psy- chiatry. 2016; 67: 59–65. https://doi.org/10.1016/j.comppsych.2016.02.013 PMID: 27095336 54. Miller AB, McLaughlin KA, Busso DS, Brueck S, Peverill M, Sheridan MA. Neural correlates of emotion regulation and adolescent suicidal ideation. Biol Psychiatry Cogn Neurosci Neuroimaging. 2018; 3: 125–132. https://doi.org/10.1016/j.bpsc.2017.08.008 PMID: 29529407 55. Lidow MS. Calcium signaling dysfunction in schizophrenia: a unifying approach. Brain Res Brain Res Rev. 2003; 43: 70–84. https://doi.org/10.1016/s0165-0173(03)00203-0 PMID: 14499463 56. Glausier JR, Lewis DA. Mapping pathologic circuitry in schizophrenia. Handbook of clinical neurology. 2018. pp. 389–417. https://doi.org/10.1016/B978-0-444-63639-3.00025-6 PMID: 29496154 PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 20 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia 57. Broadbelt K, Jones L. Evidence of altered calmodulin immunoreactivity in areas 9 and 32 of schizo- phrenic prefrontal cortex. J Psychiatr Res. 2008; 42: 612–621. https://doi.org/10.1016/j.jpsychires. 2007.07.006 PMID: 18289558 58. Saia-Cereda VM, Santana AG, Schmitt A, Falkai P, Martins-de-Souza D. The nuclear proteome of white and gray matter from schizophrenia postmortem brains. Mol Neuropsychiatry. 2017; 3: 37–52. https://doi.org/10.1159/000477299 PMID: 28879200 59. Chin D, Means AR. Calmodulin: a prototypical calcium sensor. Trends Cell Biol. 2000; 10: 322–8. https://doi.org/10.1016/s0962-8924(00)01800-6 PMID: 10884684 60. Solà C, Barrón S, Tusell JM, Serratosa J. The Ca2+/calmodulin system in neuronal hyperexcitability. Int J Biochem Cell Biol. 2001; 33: 439–55. https://doi.org/10.1016/s1357-2725(01)00030-9 PMID: 11331200 61. Sharma RK, Parameswaran S. Calmodulin-binding proteins: A journey of 40 years. Cell Calcium. 2018; 75: 89–100. https://doi.org/10.1016/j.ceca.2018.09.002 PMID: 30205293 62. Luo X-J, Li M, Huang L, Steinberg S, Mattheisen M, Liang G, et al. Convergent lines of evidence support CAMKK2 as a schizophrenia susceptibility gene. Mol Psychiatry. 2014; 19: 774–83. https://doi.org/10. 1038/mp.2013.103 PMID: 23958956 63. Takemoto-Kimura S, Suzuki K, Horigane SI, Kamijo S, Inoue M, Sakamoto M, et al. Calmodulin kinases: essential regulators in health and disease. J Neurochem. 2017; 141: 808–818. https://doi.org/ 10.1111/jnc.14020 PMID: 28295333 64. Means AR. Commentary: the year in basic science: calmodulin kinase cascades. Mol Endocrinol. 2008; 22: 2759–2765. https://doi.org/10.1210/me.2008-0312 PMID: 18845671 65. Fatemi SH, Folsom TD. The neurodevelopmental hypothesis of schizophrenia, revisited. Schizophr Bull. 2009; 35: 528–548. https://doi.org/10.1093/schbul/sbn187 PMID: 19223657 66. Zhang W, Wang D, Volk E, Bellen HJ, Hiesinger PR, Quiocho FA. V-ATPase V0 sector subunit A1 in neurons is a target of calmodulin. J Biol Chem. 2008; 283: 294–300. https://doi.org/10.1074/jbc. M708058200 PMID: 17933871 67. Schwaller B, Meyer M, Schiffmann S. “New” functions for “old” proteins: The role of the calcium-binding proteins calbindin D-28k, calretinin and parvalbumin, in cerebellar physiology. Studies with knockout mice. The Cerebellum. 2002; 1: 241–258. https://doi.org/10.1080/147342202320883551 PMID: 12879963 68. Schwaller B. The regulation of a cell’s Ca2+ signaling toolkit: The Ca2+ homeostasome. Springer, Dor- drecht; 2012. pp. 1–25. https://doi.org/10.1007/978-94-007-2888-2_1 PMID: 22453936 69. Permyakov EA, Uversky VN, Permyakov SE. Parvalbumin as a pleomorphic protein. Curr Protein Pept Sci. 2017; 18: 780–794. https://doi.org/10.2174/1389203717666161213115746 PMID: 27964700 70. Bastianelli E. Distribution of calcium-binding proteins in the cerebellum. The Cerebellum. 2003; 2: 242– 262. https://doi.org/10.1080/14734220310022289 PMID: 14964684 71. Bullock WM, Cardon K, Bustillo J, Roberts RC, Perrone-Bizzozero NI. Altered expression of genes involved in GABAergic transmission and neuromodulation of granule cell activity in the cerebellum of schizophrenia patients. Am J Psychiatry. 2008; 165: 1594–1603. https://doi.org/10.1176/appi.ajp.2008. 07121845 PMID: 18923069 72. Maloku E, Covelo IR, Hanbauer I, Guidotti A, Kadriu B, Hu Q, et al. Lower number of cerebellar Purkinje neurons in psychosis is associated with reduced reelin expression. Proc Natl Acad Sci. 2010; 107: 4407–4411. https://doi.org/10.1073/pnas.0914483107 PMID: 20150511 73. Brisch R, Bielau H, Saniotis A, Wolf R, Bogerts B, Krell D, et al. Calretinin and parvalbumin in schizo- phrenia and affective disorders: a mini-review, a perspective on the evolutionary role of calretinin in schizophrenia, and a preliminary post-mortem study of calretinin in the septal nuclei. Front Cell Neu- rosci. 2015; 9: 393. https://doi.org/10.3389/fncel.2015.00393 PMID: 26578879 74. Boley AM, Perez SM, Lodge DJ. A fundamental role for hippocampal parvalbumin in the dopamine hyperfunction associated with schizophrenia. Schizophr Res. 2014; 157: 238–243. https://doi.org/10. 1016/j.schres.2014.05.005 PMID: 24888524 75. Abdul-Monim Z, Neill JC, Reynolds GP. Sub-chronic psychotomimetic phencyclidine induces deficits in reversal learning and alterations in parvalbumin-immunoreactive expression in the rat. J Psychophar- macol. 2007; 21: 198–205. https://doi.org/10.1177/0269881107067097 PMID: 17329300 76. Cunningham MO, Hunt J, Middleton S, LeBeau FEN, Gillies MJ, Gillies MG, et al. Region-specific reduction in entorhinal gamma oscillations and parvalbumin-immunoreactive neurons in animal models of psychiatric illness. J Neurosci. 2006; 26: 2767–2776. https://doi.org/10.1523/JNEUROSCI.5054-05. 2006 PMID: 16525056 77. Piyabhan P, Tingpej P, Duansak N. Effect of pre- and post-treatment with Bacopa monnieri (Brahmi) on phencyclidine-induced disruptions in object recognition memory and cerebral calbindin, parvalbumin, PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 21 / 22 PLOS ONE Proteomic signature of cerebellum in schizophrenia and calretinin immunoreactivity in rats. Neuropsychiatr Dis Treat. 2019; 15: 1103–1117. https://doi.org/ 10.2147/NDT.S193222 PMID: 31118643 78. Servais L, Bearzatto B, Schwaller B, Dumont M, De Saedeleer C, Dan B, et al. Mono- and dual-fre- quency fast cerebellar oscillation in mice lacking parvalbumin and/or calbindin D-28k. Eur J Neurosci. 2005; 22: 861–870. https://doi.org/10.1111/j.1460-9568.2005.04275.x PMID: 16115209 79. Farrecastany M, Schwaller B, Gregory P, Barski J, Mariethoz C, Eriksson J, et al. Differences in locomo- tor behavior revealed in mice deficient for the calcium-binding proteins parvalbumin, calbindin D-28k or both. Behav Brain Res. 2007; 178: 250–261. https://doi.org/10.1016/j.bbr.2007.01.002 PMID: 17275105 80. Wöhr M, Orduz D, Gregory P, Moreno H, Khan U, Vörckel KJ, et al. Lack of parvalbumin in mice leads to behavioral deficits relevant to all human autism core symptoms and related neural morphofunctional abnormalities. Transl Psychiatry. 2015; 5: e525–e525. https://doi.org/10.1038/tp.2015.19 PMID: 25756808 81. Brown JA, Ramikie TS, Schmidt MJ, Báldi R, Garbett K, Everheart MG, et al. Inhibition of parvalbumin- expressing interneurons results in complex behavioral changes. Mol Psychiatry. 2015; 20: 1499–1507. https://doi.org/10.1038/mp.2014.192 PMID: 25623945 82. Lungu O, Barakat M, Laventure S, Debas K, Proulx S, Luck D, et al. The incidence and nature of cere- bellar findings in schizophrenia: A quantitative review of fMRI literature. Schizophr Bull. 2013; 39: 797– 806. https://doi.org/10.1093/schbul/sbr193 PMID: 22267533 83. Zhuo C, Wang C, Wang L, Guo X, Xu Q, Liu Y, et al. Altered resting-state functional connectivity of the cerebellum in schizophrenia. Brain Imaging Behav. 2018; 12: 383–389. https://doi.org/10.1007/s11682- 017-9704-0 PMID: 28293803 84. Behan ÁT, Byrne C, Dunn MJ, Cagney G, Cotter DR. Proteomic analysis of membrane microdomain- associated proteins in the dorsolateral prefrontal cortex in schizophrenia and bipolar disorder reveals alterations in LAMP, STXBP1 and BASP1 protein expression. Mol Psychiatry. 2009; 14: 601–613. https://doi.org/10.1038/mp.2008.7 PMID: 18268500 85. Martins-De-Souza D, Gattaz WF, Schmitt A, Rewerts C, Maccarrone G, Dias-Neto E, et al. Prefrontal cortex shotgun proteome analysis reveals altered calcium homeostasis and immune system imbalance in schizophrenia. Eur Arch Psychiatry Clin Neurosci. 2009; 259: 151–163. https://doi.org/10.1007/ s00406-008-0847-2 PMID: 19165527 86. Martins-de-Souza D, Gattaz WF, Schmitt A, Maccarrone G, Hunyadi-Gulyás E, Eberlin MN, et al. Prote- omic analysis of dorsolateral prefrontal cortex indicates the involvement of cytoskeleton, oligodendro- cyte, energy metabolism and new potential markers in schizophrenia. J Psychiatr Res. 2009; 43: 978– 986. https://doi.org/10.1016/j.jpsychires.2008.11.006 PMID: 19110265 87. Pinacho R, Villalmanzo N, Meana JJ, Ferrer I, Berengueras A, Haro JM, et al. Altered CSNK1E, FABP4 and NEFH protein levels in the dorsolateral prefrontal cortex in schizophrenia. Schizophr Res. 2016; 177: 88–97. https://doi.org/10.1016/j.schres.2016.04.050 PMID: 27236410 88. Velásquez E, Nogueira FCS, Velásquez I, Schmitt A, Falkai P, Domont GB, et al. Synaptosomal prote- ome of the orbitofrontal cortex from schizophrenia patients using quantitative label-free and iTRAQ- based shotgun proteomics. J Proteome Res. 2017; 16: 4481–4494. https://doi.org/10.1021/acs. jproteome.7b00422 PMID: 28949146 PLOS ONE | https://doi.org/10.1371/journal.pone.0230400 July 8, 2020 22 / 22