biomolecules Article Effects of Cannabis Use on the Protein and Lipid Profile of Olfactory Neuroepithelium Cells from Schizophrenia Patients Studied by Synchrotron-Based FTIR Spectroscopy Sergi Saladrigas-Manjón 1 , Tanja Dučić 2 , Liliana Galindo 3,4 , Cristina Fernández-Avilés 1 , Víctor Pérez 3,5 , Rafael de la Torre 1 and Patricia Robledo 1, * 1 Integrative Pharmacology and Systems Neuroscience, IMIM-Hospital del Mar Research Institute, 08003 Barcelona, Spain; sergi.saladrigas75@gmail.com (S.S.-M.); cfernandez@imim.es (C.F.-A.); rtorre2@imim.es (R.d.l.T.) 2 ALBA-CELLS Synchrotron, MIRAS Beamline, 08290 Cerdanyola del Vallès, Spain; tducic@cells.es 3 Neuropsychiatry and Addictions Institute (INAD) of Parc de Salut Mar, 08003 Barcelona, Spain; liligalindo@gmail.com (L.G.); 61155@parcdesalutmar.cat (V.P.) 4 Department of Psychiatry, University of Cambridge, Cambridge CB2 0SZ, UK 5 CIBERSAM and Department of Psychiatry, University Autonoma of Barcelona, 08193 Barcelona, Spain * Correspondence: probledo@imim.es  Received: 22 January 2020; Accepted: 15 February 2020; Published: 19 February 2020  Abstract: Schizophrenia (SCZ) is a neurodevelopmental disorder with a high genetic component, but the presence of environmental stressors can be important for its onset and progression. Cannabis use can be a major risk factor for developing SCZ. However, despite the available data on the neurobiological underpinnings of SCZ, there is an important lack of studies in human neuronal tissue and living cells addressing the effects of cannabis in SCZ patients. In this study, we analysed the most relevant bio-macromolecular constituents in olfactory neuroepithelium (ON) cells of healthy controls non-cannabis users, healthy cannabis users, SCZ patients non-cannabis users, and SCZ patients cannabis users using Synchrotron Radiation-Fourier Transform Infrared (SR-FTIR) spectrometry and microscopy. Our results revealed that SCZ patients non-cannabis users, and healthy cannabis users exhibit similar alterations in the macromolecular profile of ON cells, including disruption in lipid composition, increased lipid membrane renewal rate and lipid peroxidation, altered proteins containing more β-sheet structures, and showed an increase in DNA and histone methylation. Notably, these alterations were not observed in SCZ patients who use cannabis regularly. These data suggest a differential effect of cannabis in healthy controls and in SCZ patients in terms of the macromolecular constituents of ON cells. Keywords: schizophrenia; olfactory neuroepithelium; SR-FTIR; cannabis 1. Introduction Schizophrenia (SCZ) is a highly prevalent chronic and severe neuropsychiatric disorder. According to the World Health Organization [1], it is estimated to affect more than 21 million people worldwide. Socially, it clearly causes a familiar and economic burden, and patients’ lives are shadowed by social stereotypes and stigmas [2]. SCZ is thought to be a neurodevelopmental disorder with a high genetic component [3]. Genetic studies have suggested several potential affected genes, such as Disc1 (disrupted-in-schizophrenia 1), Nrg1 (Neuregulin-1), or DAO (Diamine Oxidase), and other regions of linkage and positional associated chromosomal abnormalities [4]. The gene Disc1 has been shown to Biomolecules 2020, 10, 329; doi:10.3390/biom10020329 www.mdpi.com/journal/biomolecules Biomolecules 2020, 10, 329 2 of 16 contribute to SCZ onset [5], and in animal models, it has been reported that Disc1 protein aggregations, predominantly in GABAergic prefrontal interneurons are related to SCZ-like symptoms [6]. In addition, many changes in lipid membrane composition and renewal rates [7,8], and lipid metabolism, such as peroxidation [9], have been described in SCZ patients’ blood serum. Although there are still important issues regarding the use of common proteomic platforms, and sample processing, the data point towards major changes in cytoskeleton and metabolic pathways. In addition, some of the most prominent changes in SCZ have been found in glial proteins [10]. Environmental factors can play a crucial role in the development of SCZ. Thus, the double hit hypothesis states that genetic factors together with the presence of one or several environmental cues are necessary for the disease to initiate its chronic course [11]. In this sense, the highly prevalent consumption of drugs of abuse in the SCZ population, particularly cannabis, could impact the onset, development, and chronicity of the disease [12]. Cannabis sativa is a recreational drug widely consumed in Europe, especially during the ages where SCZ onset usually occurs namely, late adolescence and early adulthood [13]. Chronic cannabis use alters cognitive, rewarding and affective processes [14–16] that are also critically altered in SCZ, by acting on the endocannabinoid system, and by modulating the activity of other neurotransmitters in the brain. The endocannabinoid system (ECS) consists of two major G protein-coupled receptors: the cannabinoid type 1 (CB1 R) and type 2 (CB2 R) receptors, two main endogenous ligands: anandamide (AEA) and 2-arachidonoylglycerol (2-AG) called endocannabinoids, and their synthetizing and degrading enzymes [17]. Despite the many preclinical, genetic and neuroimaging data that have been generated, the neuropathophysiological processes that lead to SCZ onset still remain unknown [18], and there are very few reports investigating the effects of cannabis use in sub-populations of SCZ patients that use cannabis regularly. Obtaining viable neurological tissue to study how biomarkers of disease are modulated by drugs such as cannabis in living patients has proven challenging. In this sense, the olfactory neuroepithelium (ON) has great potential as a surrogate model of central nervous system function to study the molecular processes involved in some neuropsychiatric disorders such as SCZ [19]. Human olfactory sensory neurons are replaced and generated by neurogenesis continuously, throughout all the adult life, from neuronal precursors located in the apical and basal membranes [20]. Consequently, the ON contains pluripotent cells that can be easily and non-invasively collected and can proliferate in vitro and differentiate into multiple cell types, including neurons and glia [21]. Therefore, the use of this pro-neuronal tissue may open a very promising window to study neuropsychiatric disorders, and to determine the impact of drugs of abuse, including cannabis. In fact, we have recently demonstrated that ON cells obtained from cannabis users exhibit changes in cannabinoid CB1 R and serotonin 2A receptor (CB1 R-5HT2A R) heteromer expression and function, which were associated with cognitive performance [22]. In this study, we investigated the macromolecular profile of proteins and lipids of ON cells using infrared spectroscopy. Imaging of human single cells based on this technique is being progressively developed for biomedical applications. The spectral signature of a cell is a mirror image of its physiological status, and its reaction to environmental cues [23]. In order to achieve single cells resolution and spectral stability, we used Synchrotron Radiation-Fourier transformed infrared (SR-FTIR) spectroscopy at MIRAS beamline at ALBA Synchrotron. This non-destructive IR spectroscopy and imaging has a ~1000-fold higher brightness than conventional global sources, thus, by improving the signal-to-noise ratio, it is very useful for the analysis of single cells [24]. Moreover, SR-FTIR is able to provide information regarding the quantity, composition, and distribution of chemical functional groups including nucleic acids, lipids and proteins in stain-free and label-free samples [25]. Therefore, we compared the IR spectral profile of ON cells derived from SCZ patients non-cannabis users (SCZ/nc), and SCZ patients cannabis users (SCZ/c) with those of healthy controls non-cannabis users (HC/nc), and healthy controls cannabis users (HC/c). Biomolecules 2020, 10, 329 3 of 16 2. Material and Methods 2.1. Subjects and Study Design A cross-sectional study was conducted on 10 SCZ patients and 9 healthy subjects. These groups were further divided into subjects who use cannabis and subjects who do not use cannabis on a regular basis: HC/nc, HC/c, SCZ/nc and SCZ/c. The study was reviewed and approved by the local institutional ethics committee: “Clinical Research Ethical Committee of the Parc de Salut Mar, Barcelona, Spain (CEIC-PSMAR), on the 17th of November 2014 (Ref. 2014/5801/l). All subjects provided a signed informed consent after a complete description of the study and procedures and were assured of the confidentiality of the data being collected. SCZ patients were recruited from the Institute of Neuropsychiatry and Addictions (INAD) in Barcelona. Eligible patients were 18 to 50 years of age with a primary psychiatric diagnosis of SCZ confirmed by the Structured Clinical Interview for DSM-IV-TR Axis I Disorders, Clinical Trial Version (SCID-CT) [26]. The exclusion criteria for SCZ patients were: (i) exhibit any comorbid axis I or II psychiatric condition; (ii) a current or previous diagnosis of substance use disorder, (iii) have received electroconvulsive therapy; (iv) present a medical condition that could affect nasal epithelia, including allergic rhinitis and other otorhinolaryngology conditions; (v) have history of severe neurological, congenital or medical condition; (vi) have allergy to the anaesthetic drug lidocaine and (vii) present severe negative or cognitive symptoms, educational level or language barrier that might prevent the subjects from possessing the necessary capacities to communicate and/or perform the evaluations tests. Healthy subjects matched by age and gender were recruited via oral or written advertising and underwent the same structured diagnostic interview to confirm eligibility. Healthy subjects fulfilled the following inclusion criteria: subjects of both sexes and age between 18 and 50 years were free of any axis I disorder in their familiar history (first- and second-degree relatives), and in the case of cannabis users, consumption of more than 5 cannabis cigarettes per week for more than 6 months. Cannabis users from both groups were instructed to interrupt cannabis use for at least 12 h before testing so as to avoid a potential confound of acute cannabis intoxication on the biochemical assessments. 2.2. Clinical and Functional Assessment Subjects first underwent an initial evaluation that consisted of a complete physical examination, including body mass index calculation (BMI = weight in kg/height in m2 ), collection of medical and psychiatric personal and family history along with sociodemographic data. All the assessments and rating scales were conducted by experienced psychiatrists and psychiatry residents, as previously described by Galindo et al. [22]. Socioeconomic status (SES) was measured by the Hollingshead–Redlich Scale [27], whose score is derived from both education attainment and an occupation prestige. SES was then calculated as ([Occupation score × 5] + [Education score × 3]), with higher scores reflecting lower SES. Functionality level and social disability were cross-sectionally measured with the Global Assessment of Functioning (GAF) disability scale from the DSM-IV manual. Substance-related disorders were assessed with the Spanish validated version of the Psychiatric Research Interview for Substance and Mental Disorders (PRISM) [28]. 2.3. Nasal Exfoliation and Cell Culture Samples of the olfactory neuroepithelium (ON) were obtained from subjects during the morning, as previously described [22,29]. Before the nasal exfoliation, all participants were evaluated and instructed not to fast that morning. After humidification of the nasal cavity and anaesthesia with Lidocaine (Xilonibsa Aerosol® 10 mg), separate sterile interdental brushes (two for each nostril) were used to obtain samples from the upper and middle turbinate. Samples were placed inside Eppendorf tubes with 250 µL of Dulbecco’s Modified Eagle Medium/ Ham F-12 (DMEM/F12) containing 10% FBS, 2% glutamine and 1% streptomycin-penicillin (GibcoBRL) at 4 ◦ C. Cell suspensions were dissociated by mechanical disaggregation. The primary cultures were grown for 3 weeks in DMEM/F12 supplemented Biomolecules 2020, 10, 329 4 of 16 with 10% FBS at 37 ◦ C and 5% CO2 before passaging into flasks (Thermo Scientific, Madrid, Spain). Primary cultures were dissociated with 0.25% trypsin (GibcoBRL), replated at 4000 cells/cm2 into 75 cm2 flasks and cultured in DMEM/F12 with 10% FBS. Cells were then expanded by passage and banked down in aliquots after harvest, followed by storage in liquid nitrogen with 20% FBS and 10% dimethyl sulfoxide (Sigma-Aldrich, Madrid, Spain). The ON cells used for the SR-FTIR spectroscopy analyses were frozen aliquots at passage 3 at a concentration of 1 × 106 cells/mL. We added 20 µL from these aliquots in the centre of a CaF2 polished window (CRYSTAN Ltd.) submerged into a well with 500 µL of medium, thus 20,000 cells were finally plated in each window. One single window was plated for the cells of each subject. Samples were placed in the incubator during 24 h at 37 ◦ C and 5% of CO2 in order to endure proper adherence of the cells. Afterwards, the medium was removed, cells were washed with saline and the windows were placed inside the refrigerator, letting them dry overnight. Finally, we placed the samples over silica gel to avoid humidity that could potentially produce interference in our measurements. CaF2 polished windows have some optical special properties that make them the most suitable surface to analyse dried cells with synchrotron infrared light transmission spectroscopy without interference [30]. 2.4. Synchrotron-Radiation Fourier Transform Infrared (SR-FTIR) Spectroscopy The SR-FTIR measurements were performed at the FTIR spectro-microscopy facility of MIRAS beamline at the synchrotron ALBA, in Cerdanyola del Vallès, Barcelona, Spain. The IR spectroscopic absorption measurements were collected in transmission mode using an IR microscope coupled to a SR-FTIR Hyperion 3000 spectrometer and microscope (Bruker, Ettlingen, Germany). The IR microscope was equipped with an automated sample stage, which could be controlled so as to define which cells had to be measured. The aperture was set to a 10 µm × 10 µm lighted by the infrared light and each spectrum was acquired after 256 scans at spectral resolution 4 cm−1 . A mercury cadmium telluride (MCT) detector cooled with liquid nitrogen was used for collecting spectra. The infrared signal from synchrotron light used during the measurements was always at an optimal level of absorption so that detection of the macromolecular changes of the sample was ensured. The IR spectroscopic measurements were collected using the 36x objective in the IR transmission. In every sample, 20 to 32 cells were measured, and two measurements in each cell. All the regions measured were selected after using the online visible light microscope. All the settings were applied using OPUS 7.5 (developed by Bruker, Ettlingen, Germany), the software package with which Hyperion 3000 works. 2.5. Statistical Analyses For the demographic categorical data (gender and tobacco use), differences between groups were assessed using two-tailed chi-squared test. Subsequently, to determine whether the different demographic, clinical, and functional data categories showed a normal distribution, Kolmogorov-Smirnov tests were first applied. Non-parametric Kruskal–Wallis test was used to assess differences in continuous variables that did not meet the assumption of a normal distribution. When significant interactions were found, individual non-parametric comparisons for each pair were performed using the non-parametric Dunn’s post-hoc test. The non-parametric Mann Whitney test was performed for differences between two groups and when normality was not assumed. Differences in variables with a normal distribution were assessed with one-way ANOVA tests and, when significant, comparison between groups was carried out by using LSD post-hoc test. Demographic, clinical and functional data were analysed with the PASW Statistics v.18 (3. SPSS Inc (2009): PASW Statistics for Windows (Version 18.0). P-values lower than 0.05 were considered statistically significant. The principal component analysis (PCA) was performed by Unscrambler X, Version 10.3, (CAMO Software, Oslo, Norway) software packages. Next, using Unscrambler X, baseline correction and unit vector normalization was applied to the spectra. Without further corrections, a PCA was performed and calculated using the Nonlinear Iterative Partial Least Squares (NIPALS) algorithm on mean centred data, without applying any weighting method. PCA analysis showed descriptive Biomolecules 2020, 10, 329 5 of 16 information regarding where to find the major differences in the SR-FTIR spectra of the samples. Differences between groups in the major peaks of the lipid (3016 cm−1 , 2956 cm−1 , 2923 cm−1 , 2871 cm−1 , 2852 cm−1 ) and the fingerprint region (1652 cm−1 , 1538 cm−1 , 1081 cm−1 , 1240 cm−1 ) were assessed using one-way ANOVA tests, and when necessary, statistical analysis was followed by the LSD post-hoc test. In addition, since, in the PCA analysis, differences in many hidden peaks were also found, two further analyses were performed. In order to avoid any human mistake in the interpretation of these results, we decided to exclude all the outliers of the lipid and protein regions, found in the PCA, from the statistical analysis. First, the Norris Gap Derivative was applied in some intervals of the spectra using Unscrambler X (CAMO Software, Oslo, Norway), where prominent differences were found in the PCA analysis (such as around 2936 cm−1 spectra region). By using this technique, the second derivative of the spectra is obtained, showing hidden peaks, which can also be tested and could, explain reliable and important information that was missing before this last smoothing. These hidden peaks were analysed between groups by using one-way ANOVA tests, followed by LSD post-hoc tests when interaction was found. All the spectroscopy data was analysed with the PASW Statistics v.18 (3. SPSS Inc (2009): PASW Statistics for Windows (Version 18.0). P-values lower than 0.05 were considered statistically significant. Second, a deconvolution of the Amide I area (1700 cm−1 to 1600 cm−1 ) of SR-FTIR spectra was carried out by using least-squares iterative curve fitting to Gaussian line shapes in OriginPro, Version 2019 software (OriginLab Corporation, Northampton, MA, USA). The deconvolution showed up specific secondary protein structure peaks that contributes to the Amide I major peak. Since secondary protein structures are not centred in just one specific wavenumber, but their contributions exist in a larger area of the spectra, the integrated area (expressed in percentage out of the total) of each secondary structure was calculated after the deconvolution, and then compared between groups using one-way ANOVA test, followed when necessary by LSD post-hoc test. These tests were carried out with the PASW Statistics v.18 (3. SPSS Inc (2009): PASW Statistics for Windows, Version 18.0, New York, NY, USA). P-values lower than 0.05 were considered statistically significant. 3. Results 3.1. Demographic, Socioeconomic, and Functional Characteristics of the Sample The demographic, socioeconomic and functional data from all subjects are shown in Table 1. Briefly, groups did not differ in age, gender, history of nasal trauma, surgery, rhinitis or BMI. Socioeconomic status was not significantly different between groups. In contrast, differences in GAF scores were observed between groups (K = 14.60; p < 0.01), with SCZ/nc (p < 0.05) and SCZ/c (p < 0.05) showing significantly lower GAF scores than HC/nc. The percentage of tobacco users, the amount of units per week, and length of use were not different between groups. In terms of cannabis use, no significant differences were observed between groups for age of first use or units per week, while a significantly longer length of use was observed in SCZ/c with respect to HC/c (U = 1.00; p < 0.05). 3.2. FTIR Spectrometry Principal Component Analyses To study the biochemical changes underlying SCZ and how cannabis use could differentially modulate these alterations we averaged SR-FTIR spectra from each group for the lipid (Figure 1A) and protein (Figure 1B) fingerprint regions. Samples from 5 subjects in each group were prepared for measurements, although cells from one subject in the HC/c group did not survive. From each subject, we selected 20 to 32 cells, and for each cell we took two independent measures: one in the cytoplasm and another in the nucleus. The final sample size was as follows: HC/nc (n = 5 subjects × 25 cells = 246 measurements); HC/c (n = 4 subjects × 32 cells = 256 measurements); SCZ/nc (n = 5 subjects × 20 cells = 198 measurements); SCZ/c (n = 5 subjects × 27 cells = 268). Biomolecules 2020, 10, 329 6 of 16 Table 1. Demographic, socioeconomic, and functional characteristics of the sample. Healthy Controls Healthy Controls (Cannabis Schizophrenia Patients Schizophrenia Patients (Non-Cannabis Users) Users) (Non-Cannabis Users) (Cannabis Users) Age (years) † 30 (27.5–36.5) 29 (22.5–41-5) 39 (26.5–46-5) 41 (36–47) Gender; male (%) 2 (40.0) 4 (100.0) 2 (40.0) 5 (100.0) BMI (kg/m2 ) † 21.8 (19.9–25.37) 20.2 (12–25.35) 30.1 (23.54–41.45) 25.35 (24.1–29.23) History of nasal trauma/surgery; 0 (0) 1 (25) 1 (20) 1 (20) n (%) History of rhinitis; n (%) 1 (20) 0 (0) 0 (0) 0 (0) Tobacco Use Users–n, (%) 5 (100) 2 (50) 3 (60) 5 (100) Units per week † 49 (22.5–122-5) 21 (14–28) 140 (105–210) 140 (87.5–210) Length of use (years) † 12 (8–16.5) 17 9 (4–16) 23 (15.5–33.5) Cannabis Use Age first use (years) ‡ n/a 17 (2.16) n/a 17.2 (3.00) Units per week † n/a 5.5 (3.25–17.5) n/a 21 (6.5–31.5) Length of use (years) † n/a 6 (1–13) n/a 22 (16–31.5) * Hollingshead-Redlich Score † 45.5 (38.75–47) 57 (49–62.5) 52 (16.5–60,5) 52 (41–56) Global Assessment of 100 (97.5–100) 100 (92.5–100) 65 (60–75) * 65 (60–65) * Functioning (GAF) † For continuous, non-normally distributed variables (†) results are reported in median (first quartile–third quartile) and provided statistics refer to the Kruskal–Wallis or Mann–Whitney tests followed by pairwise post-hoc test when appropriate. For continuous, normally distributed variables (‡) results are reported in mean (standard deviation). BMI = Body Mass Index; n/a = not applicable; * p < 0.05 vs. healthy controls. Biomolecules 2020, 10, 329 7 of 16 Biomolecules 2019, 9, x FOR PEER REVIEW 7 of 17 Figure FTIR Figure 1. Mean 1. Mean FTIR olfactory of dried of dried olfactory neuroepithelium neuroepithelium (ON) from (ON) cells cells healthy from healthy controls controls non-cannabis non-cannabis users (HC/nc, green); healthy controls cannabis users (HC/c, light blue); schizophrenic users (HC/nc, green); healthy controls cannabis users (HC/c, light blue); schizophrenic patients patients non-cannabis users (SCZ/nc, dark blue) and schizophrenic patients cannabis users (SCZ/c, non-cannabisred). users (SCZ/nc, The main dark peaks are blue)for assigned and (A)schizophrenic the lipid region patients (3050–2800cannabis cm−1), andusers (B) the(SCZ/c, protein red). The main peaks are assigned for (A) the lipid region (3050–2800 cm −1 ), and (B) the protein fingerprint fingerprint region (1800–900 cm ). Contribution of the first two components in the principal −1 component region (1800–900 −1 ). Contribution cmanalysis (PCA) for lipids of(3050–2800 cm−1)components the first two (C,D), and for in proteins (1800–1480component the principal cm−1) (E,F). analysis In C, and E, symbols represent −1 the mean values for each subject. −1 (PCA) for lipids (3050–2800 cm ) (C,D), and for proteins (1800–1480 cm ) (E,F). In C, and E, symbols represent the mean 3.3. FTIR valuesAnalysis Spectrometry for eachin subject. the Lipid Region The PCAInfor the lipid region of the spectra, we analysed the lipid region (3050 cm−1 to 2800 the amplitude cm−1 ) isofshown the most pronounced peaks at in Figure 1C. The first two (2923 and 2852 cm−1), as well as, the ratio of symmetric CH2/CH3 vibrations (2952/2871 cm−1) after components baseline correction and vector normalization (Figure 2). The peak at 2923 cm−1(PC2) of each region of the spectra accounted for 80% (PC1) and 15% of the is mainly total related to variability. For PC1 and PC2, a maximum was observed at ~2956 cm −1 corresponding to the asymmetric stretching long phospholipidic chains [31]. This band is also influenced by CH groups, which result from vibration oxidation of CH3 (ν of asym CH2 groups CH3 ),toandhydroxymethine a minimum groups, pointing at ~2920 −1a, indicating cmto possible oxidation effect [32]. differences in the CH2 asymmetric stretching vibrations (νasym CH2 ). For PC2 one peak with maxima at ~2870 cm−1 was observed, and for both PC1 and PC2 a minimum peak at ~2844 cm−1 was detected, pointing to CH3 and CH2 symmetric stretching vibrations, respectively (Figure 1D). The PCA for the Amide I and II protein regions including the ester bound (1800 cm−1 to 1480 cm−1 ) is shown in Figure 1E. The first two components of each region of the spectra accounted for 54% (PC1) and 22% (PC2) of the total variability, respectively. The PC1 of protein bands around 1655 (Amide I) and 1545 cm−1 (Amide II) areas showed a minimum at ~1655 cm−1 (assigned to α helix secondary conformation of proteins), and a maximum at ~1620 cm−1 (mostly from cross-β-sheet structures). For PC2, a maximum at ~1635 cm−1 Biomolecules 2020, 10, 329 8 of 16 (β structure) and minimum at around ~1690 cm−1 (antiparallel β structure) were observed (Figure 1F). These data reveal differences in protein secondary structure between the different groups. 3.3. FTIR Spectrometry Analysis in the Lipid Region In the lipid region of the spectra, we analysed the amplitude of the most pronounced peaks at (2923 and 2852 cm−1 ), as well as, the ratio of symmetric CH2 /CH3 vibrations (2952/2871 cm−1 ) after baseline correction and vector normalization (Figure 2). The peak at 2923 cm−1 is mainly related to long phospholipidic chains [31]. This band is also influenced by CH groups, which result from oxidation of CH2 groups to hydroxymethine groups, pointing to a possible oxidation effect [32]. Statistical analysis revealed that both HC/c (p < 0.01) and SCZ/nc (p < 0.001) showed significantly higher amplitudes than HC/nc. In addition, Biomolecules SCZ/c 2019, 9, x FOR PEER showed a significantly lower amplitude than SCZ/nc REVIEW 8 of 17(p < 0.001), and no difference in amplitude compared to HC/nc (one-way ANOVA F(3,922) = 7.347; p < 0.001) Statistical (Figure 2A). analysis most The second revealed that both HC/c prominent (p < 0.01) and contribution wasSCZ/nc at 2852 (p cm < 0.001) showed −1 (νsym CH significantly 2 ), corresponding to higher amplitudes than HC/nc. In addition, SCZ/c showed a significantly lower amplitude than the symmetric stretching vibration of CH SCZ/nc (p < 0.001), and no difference in amplitude2 , arguing that CH group may show compared to 2HC/nc (one-way ANOVA F(3,922) = a disruption in some groups. The −1 7.347;2852 cm (Figure p < 0.001) peak has 2A). Thebeen secondshown to be a reliable most prominent indicator contribution of cm was at 2852 the−1 lipidic (νsym CH character 2), of a corresponding tosince cellular compartment the symmetric stretching increases CH2 absorption vibration ofwithCH2, arguing that CHof the presence 2 group may lipid internal show amembranes, disruption in some groups. The 2852 cm−1 peak has been shown to be a reliable indicator of the such as those of the Golgi apparatus and the endoplasmic reticulum, and could also explain the lipid lipidic character of a cellular compartment since CH2 absorption increases with the presence of metabolicinternal statuslipidof membranes, a cell [33,34]. A significant increase in the amplitude of this peak was found such as those of the Golgi apparatus and the endoplasmic reticulum, and in SCZ/nc with respect to HC/nc could also explain the lipid metabolic (p < 0.05), status while of a cell SCZ/c exhibited [33,34]. A significantaincrease significantly lower amplitude in the amplitude than SCZ/nc (p < 0.01) (one-way ANOVA F(3,922) = 2.971; p < 0.05), (Figure 2B). An increased of this peak was found in SCZ/nc with respect to HC/nc (p < 0.05), while SCZ/c exhibited a ratio significantly −1 lower amplitude than SCZ/nc (p < 0.01) (one-way ANOVA F(3,922) = 2.971; p < 0.05), 2852/2871 cm has been associated to a higher−1acyl chain unsaturation level, which is known to (Figure 2B). An increased ratio 2852/2871 cm has been associated to a higher acyl chain happen with lipid peroxidation unsaturation [30]. to level, which is known A happen significant increase with lipid was observed peroxidation in HC/c [30]. A significant with respect to increase HC/nc, while a significant was observed in HC/c decrease with respectintothis HC/nc,ratio was while found indecrease a significant SCZ/c in with this respect ratio was to SCZ/nc found in (p < 0.5) SCZ/c and to HC/c < 0.001) (pwith respect(one-way to SCZ/nc (p to HC/c (p=< 5.157; < 0.5) andF(3,922) ANOVA p < 0.01) 0.001) (one-way ANOVA (Figure F(3,922) 2C). = 5.157; p < 0.01) (Figure 2C). Major Figure 2.Figure peak peak 2. Major amplitudes amplitudes related related to theCH2 to the CH2 (A)(A) 29232923 cm-1,cm-1, (B)cm-1 (B) 2852 2852andcm-1 and (C) the ratio(C) the ratio between between 2852/2871 2852/2871 cm-1 incm-1 ONincells ON from cells from healthyhealthy controls controls non-cannabis users non-cannabis users (HC/nc); (HC/nc); healthy healthy controls cannabis controls cannabis users (HC/c); schizophrenic patients non-cannabis users (SCZ/nc) and users (HC/c); schizophrenic patients non-cannabis users (SCZ/nc) and schizophrenic patients schizophrenic patients cannabis users (SCZ/c). The data are expressed as means ± S.E.M. *p < 0.05; **p cannabis