antioxidants Article Spatial and Temporal Protein Modules Signatures Associated with Alzheimer Disease in 3xTg-AD Mice Are Restored by Early Ubiquinol Supplementation Emilio Llanos-González 1,2 , Francisco J. Sancho-Bielsa 1,2 , Javier Frontiñán-Rubio 1,2 , Yoana Rabanal-Ruíz 1,2 , Sonia García-Carpintero 1,2 , Eduardo Chicano 3 , Isabel Úbeda-Banon 1,4 , Alicia Flores-Cuadrado 1,4 , Lydia Giménez-Llort 5 , Francisco Javier Alcaín 1,2 , Juan Ramón Peinado 1,2 and Mario Durán-Prado 1,2, * 1 Department of Medical Sciences, Faculty of Medicine, University of Castilla-La Mancha, 13071 Ciudad Real, Spain 2 Oxidative Stress and Neurodegeneration Group, Faculty of Medicine, Regional Centre for Biomedical Research, University of Castilla-La Mancha, 13001 Ciudad Real, Spain 3 IMIBIC, Proteomic Unit, Maimonides Biomedical Research Institute of Cordoba, Reina Sofia University Hospital, University of Cordoba, 14071 Córdoba, Spain 4 Neuroplasticity and Neurodegeneration Laboratory, Regional Centre for Biomedical Research, Ciudad Real Medical School, University of Castilla-La Mancha, 13001 Ciudad Real, Spain 5 Faculty of Medicine, Institute of Neurosciences, Universidad Autónoma de Barcelona (UAB), 08193 Barcelona, Spain * Correspondence: mario.duran@uclm.es Abstract: Despite its robust proteopathic nature, the spatiotemporal signature of disrupted protein modules in sporadic Alzheimer’s disease (AD) brains remains poorly understood. This considered Citation: Llanos-González, E.; oxidative stress contributes to AD progression and early intervention with coenzyme Q10 or its Sancho-Bielsaa, F.J.; Frontiñán-Rubio, reduced form, ubiquinol, delays the progression of the disease. Using MALDI–MSI and functional J.; Rabanal-Ruíz, Y.; bioinformatic analysis, we have developed a protocol to express how deregulated protein modules García-Carpintero, S.; Chicano, E.; arise from hippocampus and cortex in the AD mice model 3xTG-AD in an age-dependent manner. Úbeda-Banon, I.; Flores-Cuadrado, This strategy allowed us to identify which modules can be efficiently restored to a non-pathological A.; Giménez-Llort, L.; Alcaín, F.J.; condition by early intervention with ubiquinol. Indeed, an early deregulation of proteostasis-related et al. Spatial and Temporal Protein protein modules, oxidative stress and metabolism has been observed in the hippocampus of 6-month Modules Signatures Associated with mice (early AD) and the mirrored in cortical regions of 12-month mice (middle/late AD). This Alzheimer Disease in 3xTg-AD Mice observation has been validated by IHC using mouse and human brain sections, suggesting that these Are Restored by Early Ubiquinol protein modules are also affected in humans. The emergence of disrupted protein modules with AD Supplementation. Antioxidants 2023, 12, 747. https://doi.org/10.3390/ signature can be prevented by early dietary intervention with ubiquinol in the 3xTG-AD mice model. antiox12030747 Keywords: 3xTg-AD mice; Alzheimer’s disease; coenzyme Q10; MALDI-imaging; mass spectrometry; Academic Editor: Guillermo López ubiquinol Lluch Received: 12 February 2023 Revised: 16 March 2023 Accepted: 17 March 2023 1. Introduction Published: 19 March 2023 Alzheimer’s disease (AD) is the most common form of dementia, accounting for up to 75% of all cases [1]. However, despite researchers’ efforts to understand the basis of sporadic AD, the spatio-temporal deregulation of biological processes in AD brains (other than Aβ Copyright: © 2023 by the authors. deposition and formation of Tau tangles), remains unclear. From a functional point of view, Licensee MDPI, Basel, Switzerland. the overlapping of these events hinders the understanding of the highly heterogeneous This article is an open access article proteopathic nature of AD [2]. Closely linked to the proteopathy of the disease, oxidative distributed under the terms and stress emerges as a major modulator of proteostasis, especially at prodromic AD stages [3]. conditions of the Creative Commons For instance, overexpression of Aβ induces oxidative stress, which promotes accumulation Attribution (CC BY) license (https:// of Aβ in a feedback-looping cascade, leading to cognitive impairment and dementia creativecommons.org/licenses/by/ through synaptic damage and further neuronal injuries [4]. 4.0/). Antioxidants 2023, 12, 747. https://doi.org/10.3390/antiox12030747 https://www.mdpi.com/journal/antioxidants Antioxidants 2023, 12, 747 2 of 14 This considered, anti-oxidative stress therapy is a promising means of slowing down the deleterious loop between oxidative stress and proteostasis [5]. Ubiquinol, the reduced form of coenzyme Q10 (CoQ10), is a well-known antioxidant and a bioenergetic modulator that showed significant beneficial effects against in vitro Aβ internalization [6], such as decreasing brain Aβ levels and attenuating neuroinflammation [7]. Indeed, several studies have pinpointed the protective effects of CoQ10/ubiquinol at inflammation [7] and hy- poxia [8], which led to the quantification of CoQ10 levels in patients with dementia [3]. However, the full effects of CoQ10/ubiquinol on AD have not been fully elucidated. Proteomics-based strategies constitute the best approach for studying the myriad of dis- rupted processes in different pathologies [9–11]. Among the spatial proteomic technique’s variants, MALDI mass spectrometry imaging (MALDI–MSI) is gaining interest among neurodegenerative disease researchers, due to its ability to obtain spatial information data while maintaining the cytoarchitecture of the original sample [12,13]. Despite its versatility, MALDI–MSI has been scarcely employed in AD, even though distribution of analytes is essential to comprehend the extent and localization of CNS-damaging processes [14,15] and to test the efficacy of drug treatments [16]. To investigate the spatio-temporal signatures of disrupted-functional processes in AD and its modulation by ubiquinol, we studied the peptidic profile of the 3xTg-AD mice model fed a standard or ubiquinol-supplemented diet, compared to Wild Type (WT)/non- Tg mice. Using further functional analysis tools, we identified the main biological processes, protein modules, regulated to ubiquinol in the 3xTg-AD mice brain at hippocampus and cortex after 4 or 10 m of supplementation. Our results indicate a signature composed of proteostasis- and oxidative stress-related protein modules that were deregulated early (6-month mice) in the hippocampus and emerged later (12-month mice) in cortical regions. Ubiquinol prevented the deregulation and further advancement of protein modules signatures linked to AD progression in 3xTg- AD mice. Results were validated by immune-histochemistry in both mouse and human hippocampal and cortical brain sections. 2. Materials and Methods 2.1. Animals Female triple transgenic AD (3xTgAD) mice harboring PS1/M146V, APPSwe and TauP301L transgenes (obtained from the colony established in the Medical Psychology Unit, School of Medicine, Autonomous University of Barcelona [17]), and age-matched nontransgenic (Non-Tg) controls with the same genetic background (C57BL/6x129/Sv) were group housed (2–4 per cage); these elements were maintained at 22 ± 2 ◦ C and 50–70% relative humidity under a 12 h light and dark cycle. A selection of 2-month-old 3xTg-AD mice were randomly distributed according to the supplemented diet used. A standard diet (Teklad Global; Envigo, Huntingdon, UK) was fed to Vh Group, whereas a standard diet supplemented with 60 mg/100 g ubiquinol (prepared by Envigo from Teklad Global), named Ubi group. Non-Tg mice (wild type) were fed a standard diet (WT group). Animals (9 per group) received their prescribed diet and water ad libitum from 2 months until euthanasia (6 or 12 months). Animal procedures followed European (Directive 2010/63/EU) and Spanish (RO 53/2013) legislation on the protection of animals used for scientific purposes. Experiments were previously approved by the Ethical Committee of Animal Research of the University of Castilla–La Mancha (PR-2017-02-02). 2.2. Behaviourial Tests The open field and neophobia tests were performed at 6 or 12 months of age to study different behaviors and cognitive functions. For the open field test, each mouse was released in one corner of a square plexiglass arena and allowed to freely explore the environment for 15 min. Movements of the mice were tracked, and the number of rearings per mouse was quantified as a measure of vertical Antioxidants 2023, 12, 747 3 of 14 activity. For the neophobia test, mice were introduced into the center of a new standard home cage (Makrolon, 35 × 45 × 25 cm) and the number of visited corners was recorded for a period of 30 s. 2.3. Processing of Samples 2.3.1. Murine Samples Mice were anesthetized with ketamine hydrochloride (Ketolar, 1.5 mL/kg, 75 mg/kg; Parke-Oavis, Madrid, Spain) and xylazine (Xilagesic, 0.5 mL/kg, 10 mg/kg; Calier, Barcelona, Spain), and perfused with a saline solution followed by 4% w/v paraformaldehyde in 0.1 M sodium phosphate buffer (pH 7.2). Whole brains were post-fixed in 4% w/v paraformalde- hyde for 24 h, dehydrated and then embedded in paraffin blocks. Then, 5 µm sections were deparaffinized in xylene and rehydrated in a graded series of ethanol and de-ionized water. 2.3.2. Human Samples Postmortem human brain samples (N = 18) from neuropathologically diagnosed Alzheimer’s disease (AD) (N = 12) and non-Alzheimer’s disease (NAD) (N = 6) patients were used in the present study (Supplementary Table S1). Three experimental groups were established: N = 6 NAD cases, N = 6 AD (neuropathologically diagnosed as Braak tau stages III-IV) and N = 6 AD (neuropathologically diagnosed as Braak tau stages V–VI). Cases were provided by the August Pi I Sunyer Biomedical Research Institute (IDIBAPS: B.0000575), Murcia Region Network (BIOBANC–MUR: B.0000859), CIEN Foundation Tissue biobank (BTCIEN: B.0000741) and the Principality of Asturias biobank (BPA: B. 0000827), integrated into the Spanish National Biobanks Network; the cases were treated following standard operating procedures with the appropriate approval of the Ethical and Scientific Committees. All procedures described in this research were accredited by the Ethical Committee of Clinical Research of the Ciudad Real University Hospital (PID2019- 108659RB-I00). Human samples for histological procedures were processed as previously reported [18]. For cryoprotection, tissue was embedded in a phosphate-buffered solution of 2% dimethyl sulfoxide (DMSO) and 10% glycerol for 48 h; it was subsequently embedded in a phosphate-buffered solution of 2% DMSO and 20% glycerol for 48 h. The tissues were covered in 30% sucrose and cut into 50 µm coronal sections with a freezing microtome. 2.4. MALDI Mass Spectrometry Imaging (MALDI–MSI) 2.4.1. Sample Preparation Paraffined tissues were cut at 5 µm of thickness and mounted onto an indium tin oxide slide (ITO, Sigma Aldrich, Steiheim, Germany) previously coated with poly-L-lysine (P1274-25 mg, Sigma Aldrich, Steiheim, Germany) with a custom protocol (0.1 mg/mL poly L-Lys, 37 ◦ C, 1 h). Mounted samples were heated for 1 h at 65 ◦ C to increase tissue fixation on ITO slide. Sections were then deparaffined in xylene and rehydrated in a graded series of ethanol and de-ionized water. Antigen retrieval steps were performed by heating the rehydrated samples in 10 mM sodium citrate pH 6 (AD7-950-270-0500, Enzo, Farmingdale, NY, USA) at 98 ◦ C for 30 min. After cooling the samples to 40 ◦ C at room temperature, we performed two washing steps with NH4 HCO3 10 mM buffer and a drying step in vacuum for 30 min prior to tissue digestion. 2.4.2. On-Tissue Digestion and Matrix Deposition On-tissue digestion of samples was performed by spraying 4 layers of trypsin (V5111, Promega, Madison, WI, USA) 0.1 µg/mL in 25 mM NH4 HCO3 , 10% TFE, at a flow of 10 mL/min with a SunCollect (Sunchrom, Germany) sprayer. Once the on-tissue digestion was finished, slides were dried at vacuum for 30 min before matrix deposition. MALDI matrices were deposited onto tissue using a solution based on α-cyano-4-hydroxycinnamic acid and CHCA (70990, Sigma Aldrich, Steiheim, Germany) (7 mg/mL 60% ACN 0.2% TFA). The internal calibrant Glu-Fibrinopeptide (Sigma Aldrich, Steiheim, Germany) was added to the matrix solution at a final concentration of 625 fmol/mL. The final matrix Antioxidants 2023, 12, 747 4 of 14 solution was sprayed in 8 layers: first layer at 10 mL/min, second layer at 20 mL/min, third layer at 30 mL/min, and the fourth to eighth layers at 45 mL/min; all layers were at a z-axis equal to 27.05. Once completed, slides were vacuum-dried for 30 min before analysis was performed on the mass spectrometer. 2.4.3. Data Acquisition Data were acquired with an AB 5800 TOF/TOF (ABSciex, Darmstadt, Germany). A positive reflector mode was used for all samples performing double internal calibration of the spectra using a trypsin autolysis peak an 842.508 [M + H]+ and a Glu-Fibrinopetide peak at 1570.677 [M + H]+. The m/z range for all samples was defined from 650 to 1800 (80% of visualized peptides from these samples were in this range—data not shown). The following parameters were fixed as default for all samples: laser intensity at 3200, delay extraction at 450 ns and the number of shots per pixel over the sample at 150. Twenty first shots were discarded to avoid sample background noise. Deflectors’ parameters were adjusted for each sample to ensure a resolution (FWHW) > 15,000 at GluFib mass. MSI datasets were acquired using TOF/TOF Imaging Acquisition Software (ABSciex) in a fixed spatial resolution of 150 µm for all samples. All mass spectrometry data have been deposited to the ProteomeXchange Consortium via the PRIDE repository [19] with the dataset identifier PXD037763. 2.4.4. Pre-Processing Analysis of MALDI–MSI Data Spectra data were loaded into the R statistical software (R Foundation for Statistical Computing). Cardinal MSI package was used for data processing. Data dimensionality reduction was performed with a resampling method that consisted of a linear interpolation at m/z unit using the two closest m/z points at which intensities were measured. Basic tools from the R programming language were used to obtain an average resampled spectrum of each ROI for subsequent classification and to perform, when indicated, a correlation analysis. Signal intensities were extracted and uploaded to MetaboAnalyst 4.0 [20] for statistical analysis. Datasets were previously filtered by Interquartile range (IQR) and, to make the assumptions fit correctly, data logarithmic transformation and auto-scaling were performed. Univariate analyses, such as fold-change analysis and t-test, were performed to compute summary-level statistics for each feature and individual study. Partial Least Squares for Discriminant Analysis (PLS-DA) was employed as a supervised learning method to represent the separation between both groups. 2.5. Protein Identification Proteins were annotated according to the previously published and available database MaTisse [21]. Resolution of the database was adjusted to suit the features of the study; the matching protein mass was conducted individually and checked by two independent researchers. Matches with a difference of ±0.02 were excluded. All the proteins identified are described in Supplementary Tables S2–S9. 2.6. Functional Analysis and Protein–Protein Interaction Networks For the functional enrichment study, proteins expressing statistically significant differ- ences (p < 0.05) between WT and/or Ubi compared to Vh were identified and examined for relevant functional annotations using the Metascape tool (https://metascape.org/gp/ index.html (accessed on 11 February 2023)). The functional annotations used were: GO terms (GO biological process, GO molecular function, GO cellular component), Reac- tome, and KEGG pathways. Protein–protein interaction analysis was performed using the STRINGApp plugin [22] with Cytoscape 3.9.1 software (https://cytoscape.org/ (accessed on 11 February 2023)). Interaction networks were identified using the default parameters, a confidence score of 0.4, and a maximum of 5 additional interactors. The clustering plugin clusterMaker [23] was also used to generate more readable and clustered networks. The granularity parameter used was 2.5. Additionally, the String Enrichment function of the Antioxidants 2023, 12, 747 5 of 14 STRINGApp plugin was used to identify meaningful functional annotations of those inter- action networks with a minimum of 4 members. The functional annotations used were: GO terms (GO biological process, GO molecular function, GO cellular component), Reactome, and KEGG pathways. 2.7. Validation by Immunofluorescence 2.7.1. Selection Criteria for Validation The selection of markers for validation was based on the significance of their relative abundance in WT and/or Ubi compared to Vh, and their potential correlation to the etiology of AD. To assess the percent homology between the peptide sequences of the detected masses and their potential human equivalents, a sequence alignment with the Blast tool (https://blast.ncbi.nlm.nih.gov/Blast.cgi (accessed on 11 February 2023)) was performed, using the default parameters for each selected member. Those peptides whose alignments had a percentage of identity ≤90% were excluded for further validation. 2.7.2. Immuno-Histochemistry and Imaging Protocol For mice immunofluorescence, fixed, paraffin-embedded tissue was sectioned at 5 µm and deparaffined as explained above. Sections were washed in Phosphate-buffered saline (PBS) and blocked with 1% bovine serum albumin (BSA; Sigma-Aldrich A7284, St. Louis, MO, USA) and 0.5% Triton X-100 in PBS for 1 h at room temperature. The primary antibodies used are described in Supplementary Table S2. Sections were then washed and incubated with the appropriate secondary antibodies for 1 h at room temperature. To stain nuclei, brain sections were mounted in Vectashield medium with DAPI (Vector Laboratories, H-1200). A total of ten images of three different hippocampal or entorhinal cortex sections per animal were acquired on an LSM800 confocal microscope (Zeiss) using a 40× objective at room temperature. All analyses were performed in ImageJ using regions of interest on images to determine cell intensity in the respective fluorescence channels. For human immunofluorescence, hippocampus and cortex sections were chosen. Tissue antigenicity was unmasked as described previously [24]. Subsequently, tissue was immersed in blocking buffer with 10% v/v normal donkey serum (Vector Laboratories, Burlingame, CA, USA), 0.1% Triton X-100 in Tris-buffered saline (TBS 0.05 M, pH 7.6) for 60 min at RT. Sections were incubated overnight at 4 ◦ C with the following primary antibodies: Exportin-1/CRM1, ACAD9, eIF3a; (see Supplementary Table S2) in TBS with 0.3% Triton X-100. Sections were then incubated for 2 h at room temperature with Alexa- Fluor 488 (in TBS with 0.3% Triton X-100, see Supplementary Table S2), counterstained with DAPI (0.01% in TBS, Sigma Aldrich, St Louis, MO, USA) for 5 min, mounted, and coverslipped with PVA-DABCO (Sigma–Aldrich, St Louis, MO, USA). Three images of each section were acquired (n = 90) on an LSM800 confocal microscope (Zeiss) using a 63× oil objective at room temperature. All analyses were performed in ImageJ using regions of interest on images to determine cell intensity in the respective fluorescence channels. For the cresyl–violet Nissl stain, samples were incubated in 1% cresyl–violet solution for 15 min, cleared in 1% acetic acid for 1 min and rinsed in tap water for 5 min to remove excess stain. Slides were then washed in 70% ethanol for 1 min, dehydrated through 2 × 3 min changes of absolute ethanol, cleared in 2 changes of xylene for 5 min each, and coverslipped. Stereological analysis was performed using Stereo Investigator software (MBF Bioscience coupled with a Zeiss Axio Imager M2 microscope). 2.8. Statistical Analysis Data are expressed as the mean ±SEM. Statistical analyses were carried out with GraphPad Prism v.8 software (GraphPad Inc., San Diego, CA, USA). The Mann–Whitney U test was performed on experimental data. Significant differences were considered at * p < 0.05; ** p < 0.01; *** p < 0.001. Antioxidants 2023, 12, 747 6 of 14 3. Results 3.1. Behavioural Tests Supplementary Figure S1 illustrates the variables assessed in the neophobia and openfield test, which were quantified by the number of visited corners and rearings, respectively. At 6 m of age, all the groups covered a similar number of visited corners (means: WT = 8.66; Vh = 8; Ubi = 8.33), though vertical exploration activity showed distinct patterns for Ubi mice and Vh mice (Vh vs. Ubi, p < 0.0476). Differences were found between WT and Vh mice groups (WT vs. Vh, p < 0.0303). At 12 m of age, significant differences were found in neither the open-field test (means: WT = 9.5; Vh = 14; Ubi = 9.8), nor in the neophobia test (means: WT = 5; Vh = 1.75; Ubi = 2.4). 3.2. MALDI–MSI Methodology To determine the differences in spatial peptide distribution among experimental groups, specifically in hippocampal and cortical regions, MALDI–MSI was performed on paraffined coronal mice brain sections (Supplementary Figure S2). Analysis of brain sections uncovered a total of 57.501 individual m/z values; only 5000 features remained after the filtering and exclusion of poor-quality peaks according to interquartile range. Although Antioxidants 2023, 12, x FOR PEER REVIEW no major differences were observed visually in signal intensity throughout the 7 of 15 intensity map image (Figure 1A,C,E,G), the mass spectra of both regions showed a distinct peptide profile in 6 and 12 m mice (Supplementary Figures S3 and S4). A C B D WT WT Hippocampus 6 months Legend 4 Legend WT 3 WT Cortex 6 months Vh 2 Vh Ubi 1 Ubi Component 3 0 Vh Vh -1 -2 -3 -4 -5 Ubi Ubi -1 1 -1 1 E G F H WT WT Legend Hippocampus 12 months WT Legend Vh Cortex 12 months WT Ubi Vh Ubi Vh Vh Ubi Ubi Component 1 -1 1 -1 1 Figure 1. Peptide ion images and partial least-squares discriminant analysis (PLS-DA) score plots. Figure 1. Peptide ion images and partial least-squares discriminant analysis (PLS-DA) score plots. (A,C,E,G) Representative molecular images of peptide ion detected at m/z 115,459; 96,657; 1391 and (A,C,E,G) Representative 1354, respectively. The ROIsmolecular in left andimages of peptide right brain sectionsion detected at correspond m/z 115,459; and to hippocampal 96,657; 1391 cortical and 1354,respectively. regions, respectively.PLS-DA The ROIs in leftplot 3D score andanalysis right brain sections correspond of hippocampus to hippocampal 6-m (B), cortex and 6-m (D), hip- cortical pocampusregions, 12-mrespectively. PLS-DA (F), and cortex 12-m 3D(H).score plotare Results analysis shownofashippocampus mean values 6-m (B), cortex of three 6-mper replicates (D), mouse and three hippocampus 12-mdifferent (F), andmice per12-m cortex experimental groups, (H). Results WT, Vh, are shown and Ubi, as mean respectively. values of three replicates per mouse and three different mice per experimental groups, WT, Vh, and Ubi, respectively. A supervised clustering analysis based on 3D PLS–DA was further performed to as- sess individuals in each experimental group. As shown in Figure 1B,D,F,H, three sepa- rated experimental groups were observed in 6 m hippocampus and cortex, displaying three well-clustered and independent groups. However, 3D PLS-DA for 12 m mice showed overlaps, particularly among Vh and Ubi groups. Therefore, heat-map represen- Antioxidants 2023, 12, 747 7 of 14 A supervised clustering analysis based on 3D PLS–DA was further performed to assess individuals in each experimental group. As shown in Figure 1B,D,F,H, three separated experimental groups were observed in 6 m hippocampus and cortex, displaying three well-clustered and independent groups. However, 3D PLS-DA for 12 m mice showed overlaps, particularly among Vh and Ubi groups. Therefore, heat-map representations of the top 100 significant features (Supplementary Figure S5) and top 25 correlation analysis (Supplementary Figure S6) confirmed the different profiles of tendencies in response to diet-induced changes. 3.3. Functional Analysis and Protein–Protein Interaction Analysis Masses whose relative abundances were similar between the WT and Ubi groups, and different to the Vh group, were chosen for further functional analysis. Using a depurated database [21], the m/z value of each feature was translated into the name of the putative protein. In total, 478 proteins from the 188 masses studied that met the criteria, were included in the over-representation functional annotations analysis using the Metascape bioinformatics tool (Figures 2A,C and 3A,C). Figure 2. (A,C) Bar graph of the top 20 enriched terms across input gene lists, colored by p-values, in hippocampus 6 m and cortex 6 m, respectively. “Log10 (P)” is the p-value in log base 10. (B,D) Protein–protein interaction (PPI) network analysis of differentially expressed and selected genes using STRING database in hippocampus 6 m and cortex 6 m, respectively. Each module was analyzed by plug-in STRING Enrichment in Cytoscape, showing the most significative functional annotation on each case. Red-squared nodes refer to members selected for further validation. Antioxidants 2023, 12, 747 8 of 14 Figure 3. (A,C) Bar graph of the top 20 enriched terms across input gene lists, colored by p-values, in hippocampus 12 m and cortex 12 m, respectively. “Log10 (P)” is the p-value in log base 10. (B,D) Protein–protein interaction (PPI) network analysis of differentially expressed and selected genes using STRING database in hippocampus 12 m and cortex 12 m, respectively. Each module was analyzed by plug-in STRING Enrichment in Cytoscape, showing the most significant functional annotation on each case. Red-squared nodes refer to members selected for further validation. For 6 m hippocampus, the most significant top-three functional terms were supramolec- ular fiber organization (21 proteins); translation (15 proteins), and neutrophil degranulation (17 proteins) (Figure 2A); for 6 m cortex, the most significant top-three functional terms were structural molecule activity (12 proteins), actomyosin structure organization (6 pro- teins), and cadherin binding (7 proteins) (Figure 2C). In contrast, 12 m hippocampus was mainly enriched in proteins involved in actin filament-binding (6 proteins); RHO GTPase effectors (5 proteins), and focal adhesion (4 proteins) (Figure 3A); proteins related to struc- tural molecule activity (25 proteins), cell adhesion molecule binding (19 proteins) and axon guidance (13 proteins) were enriched in 12 m cortex (Figure 3C). Many proteins identified shared similar biological functions or signaling pathways, suggesting a functional relevance. For a thorough analysis, a STRING protein–protein interactions (PPIs) network was performed based on close- functional associations or phys- ical interactions (Figures 2B,D and 3B,D). In 6 m hippocampus, this strategy revealed a significant enrichment (p = 1.0 × 10−16 ) of PPIs (edges) among the 128 proteins included (Figure 2B). Most of the PPIs were focused on six main protein modules; RNA binding (e.g., RPL26, EEF2), redox processes (e.g., CAT, PRDX1), and assemble of myofibril (e.g., EPPK1, MYH11) were the most represented. However, one cluster related to cell–substrate junctions comprising 22 edges out of 40 proteins included could be found in 6 m cortex Antioxidants 2023, 12, 747 9 of 14 (p = 1.03 × 10−10 ) (Figure 2D). A lower but significative enrichment (p = 0.0328) was re- ported from 29 proteins included in 12-m hippocampus (Figure 3B), wherein two clusters involved in actin binding (e.g., FLNA, FHL1; 13 nodes) and structure of ribosomes (e.g., RPL13A, RPS11; 4 nodes) were highlighted. As in 6 m hippocampus, a significant enrich- ment in 12 m cortical region (p = 5.73 × 10−10 ) of PPIs was observed (Figure 3D).10Out Antioxidants 2023, 12, x FOR PEER REVIEW of 15 of 82 proteins included, seven clusters were identified and the most represented were those involved in cell morphogenesis (e.g., CA2, TXN; 20 nodes), protein translation (e.g., EIF3A, RPL15; 9 nodes), 3.4. Protein andValidation Modules fatty acidbybeta-oxidation (e.g., ACSL3, ECH1; 8 nodes). Immunofluorescence We also selected a set of markers from mice brain sections (see Material and Methods) 3.4. Protein Modules Validation by Immunofluorescence (Supplementary Table S11). For each region of study, two representative proteins are We also shown. At 6 selected a set of markers m hippocampus, from mice the expression brain sections of COPG1 (see Material and HADHA and proteins Methods) was exam- (Supplementary Table S11). For each region of study, two representative proteins are ined, given their important role in protein transport and fatty acid oxidation, respectively, shown. At 6 m hippocampus, the expression of COPG1 and HADHA proteins was examined, (Figure 4A,B). Both COPG1 and HADHA significantly increased in Vh dentate gyrus com- given their paredimportant to Ubi. Inrole6 in m protein corticaltransport andexpression region, the fatty acid oxidation, respectively, of PHB2 and RPL23A (Figure 4A,B). proteins, in- Both COPG1 and HADHA significantly increased in Vh dentate gyrus compared volved in mitophagy and ribosome structure, was analyzed (Figure 4C,D). A significant to Ubi. In 6reduction m corticalinregion, the expression of PHB2 and RPL23A proteins, involved cortex for both PHB2 and RPL23A markers was also evident in Ubi group in mitophagy and ribosome compared structure, was analyzed (Figure 4C,D). A significant reduction in cortex for to Vh. both PHB2 and RPL23A markers was also evident in Ubi group compared to Vh. Hippocampus Cortex WT Vh Ubi WT Vh Ubi A C Corrected fluorescence intensity Corrected fluorescence intensity 6×10 4 1.5×10 5 WT ✱✱✱✱ ✱✱ ✱✱✱✱ ✱✱✱✱ DAPI DAPI DAPI DAPI DAPI DAPI Copg1 4×10 4 1×10 5 Phb2 Copg1 Copg1 Copg1 Phb2 Phb2 Phb2 2×10 4 5×10 4 Merge Merge Merge 0 Merge Merge Merge 0 WT Vh Ubi WT Vh Ubi B D Corrected fluorescence intensity Corrected fluorescence intensity 3×10 5 ✱✱✱✱ 2.5×10 4 ✱✱✱✱ ✱✱ ✱✱ DAPI DAPI DAPI DAPI DAPI DAPI 2×10 4 Hadha Rpl23a 2×10 5 1.5×10 4 Hadha Hadha Hadha Rpl23a Rpl23a Rpl23a 1×10 4 1×10 5 5×10 3 Merge Merge Merge Merge Merge Merge 0 0 WT Vh Ubi WT Vh Ubi WT Vh Ubi WT Vh Ubi E G Corrected fluorescence intensity Corrected fluorescence intensity 4×10 5 1.5×10 5 ✱✱✱✱ ✱✱✱✱ ✱✱ ✱✱✱✱ DAPI DAPI DAPI DAPI DAPI DAPI 3×10 5 Rpl15 Xpo1 1×10 5 2×10 5 Xpo1 Xpo1 Xpo1 Rpl15 Rpl15 Rpl15 5×10 4 1×10 5 Merge Merge Merge Merge Merge Merge 0 0 F WT Vh Ubi H WT Vh Ubi Corrected fluorescence intensity 1.5×10 5 Corrected fluorescence intensity 2×10 5 ✱✱✱ ✱✱ ✱✱✱✱ DAPI DAPI DAPI DAPI DAPI DAPI ✱✱✱ 1.5×10 5 Acad9 1×10 5 Eif3a 1×10 5 Acad9 Acad9 Acad9 Eif3a Eif3a Eif3a 5×10 4 5×10 4 Merge Merge Merge Merge Merge Merge 0 0 WT Vh Ubi WT Vh Ubi Figure4.4.Protein Figure Proteinexpression expression ofof selected selected markers markers inin mice mice sections. sections. (A–H) (A–H) Representative Representative confocal confocal imagesof images of selected selected markers markersin inthe thehippocampus hippocampus and andcortical regions cortical of 6of regions m 6and m 12 m 12 and oldmbrain old mice brain sections. mice Scale Scale sections. bars: 200 μm. bars: Graphs 200 show mean µm. Graphs show values mean± values S.E.M. of ± corrected S.E.M. offluorescence intensities corrected fluorescence from 10 measurements from section, three sections per mice and three different mice for each ex- intensities from 10 measurements from section, three sections per mice and three different mice for perimental group, WT, Vh and Ubi, respectively. Significant differences were determined using the each experimental two-tailed group, WT, Mann–Whitney test Vh (*p