Ramiro et al. BMC Neuroscience (2023) 24:7 BMC Neuroscience https://doi.org/10.1186/s12868-023-00775-7 RESEARCH Open Access Influence of sex, age and diabetes on brain transcriptome and proteome modifications following cerebral ischemia Laura Ramiro1, Júlia Faura1, Alba Simats1, Paula García‑Rodríguez1, Feifei Ma1, Luna Martín2, Francesc Canals2, Anna Rosell1 and Joan Montaner1,3* Abstract Ischemic stroke is a major cause of death and disability worldwide. Translation into the clinical setting of neuropro‑ tective agents showing promising results in pre-clinical studies has systematically failed. One possible explanation is that the animal models used to test neuroprotectants do not properly represent the population affected by stroke, as most of the pre-clinical studies are performed in healthy young male mice. Therefore, we aimed to determine if the response to cerebral ischemia differed depending on age, sex and the presence of comorbidities. Thus, we explored proteomic and transcriptomic changes triggered during the hyperacute phase of cerebral ischemia (by transient intraluminal middle cerebral artery occlusion) in the brain of: (1) young male mice, (2) young female mice, (3) aged male mice and (4) diabetic young male mice. Moreover, we compared each group’s proteomic and transcriptomic changes using an integrative enrichment pathways analysis to disclose key common and exclusive altered proteins, genes and pathways in the first stages of the disease. We found 61 differentially expressed genes (DEG) in male mice, 77 in females, 699 in diabetics and 24 in aged mice. Of these, only 14 were commonly dysregulated in all groups. The enrichment pathways analysis revealed that the inflammatory response was the biological process with more DEG in all groups, followed by hemopoiesis. Our findings indicate that the response to cerebral ischemia regarding prot‑ eomic and transcriptomic changes differs depending on sex, age and comorbidities, highlighting the importance of incorporating animals with different phenotypes in future stroke research studies. Keywords Ischemic stroke, Proteome, Transcriptome, Sex, Diabetes, Age Introduction Ischemic stroke is a leading cause of death and disability worldwide [1]. Nowadays, recombinant tissue-plasmino- gen activator (rt-PA) and tenecteplase (TNK) are the only *Correspondence: approved drug to treat acute ischemic stroke, although Joan Montaner joan.montaner@vhir.org cerebral blood flow can also be restored by removing 1 Neurovascular Research Laboratory, Vall d’Hebron Institute of Research the obstructive clot through mechanical thrombectomy (VHIR), Universitat Autònoma de Barcelona, Barcelona, Spain 2 in selective patients [2, 3]. Despite the effectiveness of Proteomics Laboratory, Vall d’Hebron Institute of Oncology (VHIO), Universitat Autònoma de Barcelona, Barcelona, Spain these reperfusion strategies, their short therapeutic win- 3 Stroke Research Program, Institute of Biomedicine of Seville, IBiS/ dow, some potential side effects and the strict clinical Hospital Universitario Virgen del Rocío/CSIC/University of Seville & eligibility criteria impede offering these treatments to all Department of Neurology, Hospital Universitario Virgen Macarena, Seville, Spain stroke patients [4, 5]. Neuroprotective agents have been proposed as adjuvant therapies to be given during the © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​ mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Ramiro et al. BMC Neuroscience (2023) 24:7 Page 2 of 15 hyperacute phase to minimize the impaired brain-dam- and exclusive dysregulated proteins, genes and pathways aged area while the ischemic brain is waiting for reperfu- playing a crucial role in the first stages of the disease, sion therapy to restore the blood flow [6]. To date, more with the hypothesis that those molecular changes could than 1,000 neuroprotective agents have shown promising further determine the response to a given drug. results in preclinical models [7], but translation into the clinical setting has systematically failed [7–9]. Materials and methods The failure of the clinical trials with neuroprotective Study design agents for the acute treatment of ischemic stroke makes The study presented here consists of a discovery study the scientific community reconsider the most classic performed on RNA and protein extracts from mice designs of preclinical studies. One of the reasons that ischemic brains of four different groups: (1) young male could explain this failure in translation from “bench to mice, (2) young female mice, (3) aged male mice and (4) bedside” is that the animal models used to test neuro- diabetic young male mice, using mass spectrometry and protectants do not represent properly the demographic RNA microarrays, with further integrative bioinformat- and health characteristics of stroke patients [10]. Nor- ics analysis. Each group included 8 ischemic animals, and mally, pre-clinical studies include young animals, while it 4 sham-control animals were used to discard genes and is known that more than 65% of all strokes occur in sub- proteins differentially altered due to other phenomena jects older than 65 years [11] and only 10% in individuals rather than the ischemic event. Of note, the analysis of under 50 years old [12, 13]. Regarding sex, most pre-clin- young male mice was performed in a first set of experi- ical studies are performed on male animals, while stroke ments [20]. The other 3 groups of animals were per- greatly impacts both sexes [14], being the first cause of formed in a second set of experiments simultaneously. death in women in some European countries [13, 15]. Finally, the vast majority of pre-clinical studies are per- Animals formed in healthy animals, although stroke patients usu- All animal procedures were performed in compliance ally present a high incidence of comorbid conditions with the Spanish legislation and in accordance with the such as hypertension, diabetes and/or obesity, which are Directives of the European Union and were approved known to be among the most common stroke risk factors by the Ethics Committee of the Vall d’Hebron Research [16]. In this regard, the Stroke Therapy Academic Indus- Institute (protocol numbers 03/19 and 73/20). All experi- try Roundtable (STAIR) published some recommenda- ments were conducted in a randomized manner and in tions to improve the quality of preclinical research and adherence to the ARRIVE guidelines [21]. Group descrip- to make clinical implementation more efficient. Of note, tion: (1) C57BL/6 J male mice 8/12-week-old (referred these recommendations give special attention to spe- as males), (2) C57BL/6 J female mice 8/12-week-old cies, strain and experimental model selection. In brief, (referred as females), (3) C57BL/6 J male mice 18-month- they highlight the necessity of using both sexes in pre- old (referred as old [22]) and (4) BKS-Leprdb/db/JOrlRj clinical studies and highly recommend including animals diabetic male mice 8/12-week-old (referred as diabetics) with comorbidities and aged animals to better represent (Janvier Labs, France). Animals were kept in a climate- the human population that most likely will suffer stroke controlled environment on a 12-h light/12-h dark cycle. [17–19]. Food and water were available ad libitum. Analgesia In this regard, it is plausible to think that the response (Buprenorfine, 0.05 mg/kg, s.c, Divasa Farma-Vic S.A, to cerebral ischemia might be different depending on Spain) was administered to all animals to minimize pain sex, age and comorbidities, and understanding these and discomfort. Anesthesia (isoflurane, 4% for induction, differences at the molecular level in the brain is funda- 2% for maintenance in medicinal air, Abbot Laboratories, mental to better comprehend stroke pathophysiology Spain) was given via facemask during all surgical proce- and to improve the “bench to bedside” translation. With dures described below. this background, we aimed to explore proteomic and A total of 57 animals were used for the present study. transcriptomic changes at a brain level triggered dur- Among them, 9 were excluded after applying the follow- ing the hyperacute phase of cerebral ischemia in mice ing criteria: incomplete occlusion or reperfusion after of four different groups: (1) young male mice, (2) young removal of the filament (n = 2) and death during the female mice, (3) aged male mice and (4) diabetic young experimental protocol (n = 7), while the other 48 were male mice. We aimed to compare the proteomic and the included. transcriptomic changes of each group using an integra- tive enrichment pathways analysis to determine if the Transient middle cerebral artery occlusion (MCAO) model response to ischemic stroke differed between the stud- Transient ischemia in the territory of the middle cer- ied groups. Moreover, we aimed to reveal key common ebral artery (MCA) was induced by introducing an Ramiro et al. BMC Neuroscience (2023) 24:7 Page 3 of 15 intraluminal filament through the external carotid artery, Genechip© Mouse Clariom S 24 × arrays plates (Affyme- as described before [23]. Briefly, animals were anesthe- trix, ThermoFisher). Starting material was 100 ng of total tized and body temperature was maintained at 37 °C RNA of each sample. Briefly, sense ssDNA was generated using a heating pad. The regional cerebral blood flow from total RNA with the GeneChip WT Plus Reagent Kit (CBF) was monitored close to the region irrigated by the (Affymetrix) following the manufacturer’s instructions. MCA during the whole process by affixing a laser Dop- Then, sense ssDNA was fragmented, labeled and hybrid- pler probe (Moor Instruments, UK) to the skull. After- ized to the arrays with the GeneChip WT Terminal Labe- ward, animals were placed in the supine position and ling and Hybridization Kit (Affymetrix). Arrays plates the right bifurcation of the external carotid artery and were scanned and processed with Affymetrix GeneChip internal carotid artery were exposed. Then, a silicone- Command Console to obtain expression array intensity. coated nylon monofilament (Doccol Corporation, USA; cel files. 602256PK10for males and females, 702256PK5 for dia- betics and 602456PK10 for old animals) was introduced Proteomics study through the external carotid artery to occlude the MCA. Due to equipment availability, the proteomics study was MCA occlusion (MCAO) was confirmed by a reduction performed following two different protocols depending in the cortical CBF recorded by the laser Doppler probe. on the group of mice. Young male mice proteomic study Next, the incision was closed with a silk suture and ani- was previously published [20]. Briefly, after in-gel diges- mals were allowed to recover from anesthesia. Ninety tion of the proteins, tryptic digests were analyzed using minutes later, mice were re-anesthetized and the filament an LC–MS approach in a linear trap quadrupole (LTQ) was removed to induce reperfusion of the MCA. After Orbitrap Velos mass spectrometer (ThermoFisher). reperfusion, mice were allowed to recover from anes- For the remaining groups of animals, previous to thesia under supervision. To ensure the correct occlu- LC-MS analysis, samples were digested with trypsin sion of the MCA, only animals that exhibited a reduction in-solution. Initial buffer exchange to 8 M Urea 50 mM of ≥ 75% of CBF after filament introduction and a recov- ammonium bicarbonate (AB) was performed using ery of ≥ 75% after filament removal were included in the 0.5 mL 3KDa cut-off Amicon Ultra ultrafiltration devices study [24]. Sham animals underwent the same surgical (Merck-Millipore). Total protein content was quantified procedures without the insertion of the nylon filament, using an RCDC kit (Bio-Rad), and 15 μg of each protein and therefore, without MCAO. extract was taken for tryptic digestion. Samples were first reduced with DTT to a final concentration of 10 mM, for Brain collection and extraction of protein and RNA 1 h at room temperature (RT), and then alkylated with Mice were deeply anesthetized (5% isoflurane) 30 min 20 mM of iodoacetamide (IAA) for 30 min at rt in the after reperfusion and transcardially perfused with 20 mL dark. Carbamidomethylation reaction was quenched by of cold saline. Immediately after perfusion, mouse brains the addition of N-acetyl-L-cysteine to a final concentra- were quickly removed and sectioned into 6 slices of tion of 35 mM followed by incubation for 15 min at RT 1 mm in cold conditions. The slice corresponding to the in the dark. Samples were diluted with 50 mM AB to a bregma anatomical point (representing the core of the final concentration of 1 M Urea, and then modified por- infarct [25]) was carefully dissected to separate the right cine trypsin (Promega Gold) was added in a ratio of 1:20 (ipsilateral, IP) and left (contralateral, CL) hemispheres (w/w), and the mixture was incubated overnight at 37 °C. separately. Each hemisphere was flash-frozen in liquid The reaction was stopped with formic acid (FA) to a final nitrogen and stored at − 80 °C. concentration of 0.5%. The tryptic digests were then puri- Flash-frozen tissues were pulverized into powder in liq- fied on SCX spin columns (PolyLC), evaporated to dry- ­ irVana™ ­Paris™ kit (Thermo uid nitrogen, and total fractions of protein and RNA were ness and stored at -20ºC until analyzed. then isolated using the M For LC-MS/MS analysis tryptic digests were diluted in Fisher Scientific Inc., USA) following manufacturers’ 3% ACN, 1% FA and 500 ng of the sample was loaded to instructions. RNA and protein fractions were stored at a 300 μm × 5 mm Pep-Map C18 (Thermo Scientific) at a − 80ºC until further use. flow rate of 15 μl/min using a Thermo Scientific Dionex Ultimate 3000 chromatographic system (Thermo Scien- Transcriptomics study tific). Peptides were separated using a C18 analytical col- Total RNA concentrations from brain samples were umn (NanoEase MZ HSS T3 column, 75 μm × 250 mm, measured with a Nanodrop 1000 Spectrophotometer 1.8 μm, 100 Å, Waters) with a 210 min run, comprising (ThermoFisher) and RNA integrity was assessed using four consecutive steps, first 3 min of isocratic gradient at the Agilent 2100 BioAnalyzer (Agilent Technologies, 3%B, from 3 to 35% B in 180 min, from 35 to 50% B in USA). Gene expression patterns were analyzed using 5 min, from 50 to 85% B in 1 min, followed by isocratic Ramiro et al. BMC Neuroscience (2023) 24:7 Page 4 of 15 elution at 85% B in 5 min and stabilization to initial con- Bioinformatics and statistical analyses ditions (A = 0.1% FA in water, B = 0.1% FA in CH3CN). Statistical and bioinformatics analyses of data were The column outlet was directly connected to an Advion performed using custom scripts in R language (version 4.0.2) and R Studio (R Core Team, 2017, Vienna, Aus- Fusion ­Lumos™ Tribrid (Thermo Scientific). The mass TriVersa NanoMate (Advion) fitted on an Orbitrap tria) with common Bioconductor packages. spectrometer was operated in data-dependent acquisi- As the male transcriptomic and proteomic studies tion (DDA) mode. Survey MS scans were acquired in the were carried out in a first set of experiments separately orbitrap with the resolution (defined at 200 m/z) set to from the other 3 groups, all groups have been analyzed 120,000. The lock mass was user-defined at 445.12 m/z independently. in each Orbitrap scan. The top-speed (most intense) ions per scan were fragmented in the linear ion trap (CID) and detected in the Ion Trap. Quadrupole isolation was Transcriptomics data employed to selectively isolate peptides of 350–1700 m/z. Transcriptomics data analyses have been performed The predictive automatic gain control (pAGC) target separately for each group of animals. R olgo package was set to 1.5e5. The maximum injection time was set to was used to upload.cel files. Quality control of the data 50 ms for MS1 and 35 ms for MS2 scan. Included charged was performed using the arrayQualityMetrics package states were 2–7. Target ions already selected for MS/MS [26], a principal component analysis (PCA) and inten- were dynamically excluded for 30 s. The mass tolerance sity boxplots for each file. Afterward, Robust Multiar- of this dynamic exclusion was set to ± 2.5 ppm from the ray Average algorithm [27] was used for pre-processing calculated monoisotopic mass. The spray voltage in the microarray data to perform background adjustment, NanoMate source was set to 1.7 kV. RF Lens were tuned log2 transformation and quantile normalization. To to 30%. The minimal signal required to trigger MS to MS/ obtain the expression measure of each probe, a lin- MS switch was set to 5000 and activation Q was 0.250. ear model was fit to the normalized data. Next, genes The spectrometer was working in positive polarity mode whose standard deviation (SD) was below the 75 per- and singly charge state precursors were rejected for centile of all SD were filtered out from the whole Progenesis® QI for proteomics software v 3.0 (Nonlin- fragmentation. dataset. The selection of differentially expressed genes was ear dynamics, UK) was used for MS data analysis using based on a linear model analysis with empirical Bayes default settings. The results from LCMS runs were auto- modification for the variance estimates [28], using the matically aligned to a selected reference sample. Align- limma package. This analysis has been performed consid- ments were then manually supervised. Normalization of ering sample pairing. False Discovery Rate (FDR)-based MS signals, based on the median of the ratiometric distri- corrections for multiple testing were also calculated [29]. ­ rogenesis® QI for proteomics soft- bution of the abundance measurements, was performed Fold-change (FC) was calculated dividing the IP by the automatically by the P CL expression value for each animal. ware. Only features within the 400 to 1600 m/z range, Sham animals’ data were used to discard genes differ- from 50 to 180 min of retention time, and with positive entially expressed due to other phenomena rather than charges between 2 to 4 were considered for identifica- cerebral ischemia. With this purpose, genes differen- ­ rogenesis® QI for proteomics and searched tion and quantification. Peaklists (mgf files) were gener- tially expressed when comparing IP and CL both in sham ated using P and MCAO animals (FDR < 0.25) with a difference in from Proteome Discoverer 2.1 (Thermo Fisher Scientific logFC < 0.5 were discarded. Inc., USA) using the Mascot search engine (v5.1, Matrix Science, UK). Protein identification was carried out using the SwissProt-MusMusculus database (2020, Jul Proteomics data 01: 55,400 entries) setting precursor mass tolerance to Protein abundance quantification was based on the sum 10 ppm and fragment mass tolerance to 0.5 Da. Oxidized of the peak areas within the isotope boundaries of pep- methionine was considered as a variable amino acid tide ion peaks. Proteins identified by identical peptide modification and carbamidomethylation of cysteines as sets were grouped to satisfy the principles of parsimony. a fixed modification. Trypsin was selected as the enzyme Only those proteins quantified and identified with at allowing up to two missed cleavages. The significance least 2 unique and non-conflicting peptides (it is, fea- threshold for the identifications was set to p < 0.05, mini- tures assigned unambiguously to peptides belonging to mum Mascot ions score of 20. Only those proteins quan- the protein, as assessed by Progenesis software). Protein tified and identified with at least 2 unique peptides were abundance values were log-10 transformed and column- considered for further statistical analyses. wise standardized [30]. Quality control of the data was Ramiro et al. BMC Neuroscience (2023) 24:7 Page 5 of 15 performed using a principal component analysis (PCA) in PCA with lower intensities than the other individu- and intensity boxplots for each file. als, which remained after normalization. For that reason, The selection of differentially expressed proteins was these 2 animals were removed from the analysis. After- based on a linear model analysis with empirical Bayes ward, all samples passed the quality control. So, for the modification for the variance estimates [28], using the transcriptomics study young female mice, aged male limma package. This analysis has been performed consid- mice and diabetic young male mice groups included 8 ering sample pairing. False Discovery Rate (FDR)-based ischemic animals and 4 sham-control animals, while the corrections for multiple testing were also calculated [29]. young male mice group included 6 ischemic animals and Fold-change (FC) was calculated dividing the IP by the 4 sham-control animals. CL expression value for each animal. Differentially expressed genes Integrative analyses In male mice, we found 61 differentially expressed genes An integrative analysis of the transcriptomics data of (DEG) after MCAO when comparing the IP and the CL MCAO male, female, old and diabetic mice has been car- brain regions (FDR < 0.25), while in sham-control ani- ried out. Venn diagrams were used to represent common mals we did not identify any DEG. In female mice, there and exclusive differentially expressed genes of studied were 77 DEG between IP and CL brain regions in MCAO groups. The common genes were used independently animals (FDR < 0.25), while only 8 DEG were found in as input for STRING protein-protein interaction (PPI) sham animals (FDR < 0.25) and there were no common network analysis. The interactions were based on text dysregulated genes between female MCAO and sham mining, experiments, databases, co-expression, neigh- animals. In diabetic mice, 699 DEG were found when borhood, gene fusion and co-occurrence, considering a comparing the IP and the CL brain regions in MCAO medium confidence (0.4) [31]. animals (FDR < 0.25), while in sham control animals there For the integrative analysis, ActivePathways and its were no DEG. Finally, in old mice we found 24 DEG were corresponding R package were used. In brief, this tool found when comparing the IP and the CL brain regions merges p-values and performs ranked hypergeometric in MCAO animals (FDR < 0.25), while in sham control tests to determine enriched pathways and processes [32]. animals there were no DEG. Details of the top 15 DEG First, a matrix was created with all the analyzed genes after cerebral ischemia of each studied group are shown and their corresponding p-values for each dataset. Then, in Table 1. the p-values were merged using Brown’s method [33], obtaining a ranked gene list. Finally, ranked hypergeo- Integrative analysis of transcriptomic results metric tests were carried out to determine if a pathway is To understand the mechanisms underlying ischemic enriched in the ranked gene list. The following functional stroke, we sought to identify common DEG after cerebral gene sets were used: Gene Ontology (GO) Biological Pro- ischemia in all groups, as they might be playing a cru- cesses and GO Molecular Functions [34, 35]. cial role in ischemic stroke response and thus be point- To visualize the enriched pathways as a network, ing at crucial therapeutic targets in the acute phase of Cytoscape software [36] and the EnrichmentMap app the disease. We found 14 DEG in the ischemic brain that [37] were used. were common in all four groups of animals (IP vs CL, FDR < 0.25). In addition, it is worth noting that 27 genes Results were dysregulated only in female mice, 640 in diabetic Transcriptomics results mice, 4 in old mice and 16 in male mice (Fig. 1, Addi- We determined changes in gene expression that occur tional file 1: Table S1). in the brain within the hyperacute phase after ischemic The multi-omics pathway enrichment analysis was stroke by analyzing mouse-brain samples obtained 2 h conducted on the integrated gene data to further under- after cerebral ischemia induction (30 min after reperfu- stand transcriptomic alterations triggered by cerebral sion) or sham surgery. The changes in gene expression ischemia. Interestingly, this analysis pointed out enriched were elucidated using microarrays, comparing the IP pathways on every single dataset but also enriched path- and CL regions and correcting by sham animals for each ways that are only found through data integration that studied group. are not apparent in any single omics dataset indepen- Quality control of samples was performed for each dently [32]. This analysis identified several biological group of animals. Regarding diabetic, female and old processes whose genes are significantly enriched within mice all samples passed the quality control. However, the hyperacute phase of cerebral ischemia. The top 20 after performing quality control in male mice, 2 ischemic enriched biological processes (sorted by p-value) and the animals (4 samples) were clearly separated from the rest number of DEG for each group are shown in Fig. 2A. We Ramiro et al. BMC Neuroscience Table 1 Top 15 differentially expressed genes (FDR < 0.25) between the infarcted hemisphere and the contralateral healthy hemisphere 2 h after cerebral ischemia induction. Genes that are differentially expressed in all the groups of animals are highlighted in bold. FDR: false discovery rate; logFC: Logarithmic fold change (2023) 24:7 Young male mice, males Young female mice, females Diabetic young male mice, diabetic Aged male mice, old SYMBOL logFC P.Value FDR SYMBOL logFC P.Value FDR SYMBOL logFC P.Value FDR SYMBOL logFC P.Value FDR CCL3 2.4958 1.73E-15 8.80E-12 ATF3 1.4557 1.42E-06 0.00284 HSPA1A 2.4544 1.02E-09 5.17E-06 NPAS4 2.9274 1.53E-08 7.75E-05 FOSB 1.9774 7.77E-12 1.58E-08 PTGS2 1.5976 1.76E-06 0.00284 FOS 2.4266 9.29E-09 1.63E-05 CCL3 2.0499 1.08E-07 2.74E-04 CCN1 1.8493 1.10E-11 1.58E-08 CCRL2 1.1296 2.21E-06 0.00284 CCN1 1.9876 9.61E-09 1.63E-05 FOS 1.6891 8.48E-07 0.00143 FOS 2.1172 1.25E-11 1.58E-08 FOS 1.5602 2.24E-06 0.00284 CCL3 2.6185 1.55E-08 1.97E-05 IL1A 1.7778 2.60E-06 0.00329 NPAS4 2.6951 7.08E-11 7.18E-08 CCL3 2.1777 2.93E-06 0.00298 FOSB 2.1383 4.27E-08 4.33E-05 PTGS2 1.3054 5.90E-06 0.00599 ATF3 1.6062 1.09E-09 9.22E-07 NPAS4 2.2737 4.40E-06 0.00372 GADD45G 1.7414 5.81E-08 4.91E-05 FOSB 1.5304 9.64E-06 0.00815 JUNB 1.4066 2.36E-09 1.67E-06 ADAMTS1 1.0314 6.65E-06 0.00482 IER2 1.2020 9.91E-08 7.18E-05 DUSP6 0.9922 3.29E-05 0.02382 RGS2 1.4974 2.80E-09 1.67E-06 CCN1 1.3156 9.48E-06 0.00601 THBS1 1.8039 1.14E-07 7.22E-05 HSPA1A 1.6433 7.39E-05 0.04390 GADD45G 1.3902 2.96E-09 1.67E-06 NR4A1 1.0643 1.54E-05 0.00868 NPAS4 3.1196 3.98E-07 2.25E-04 JUN 0.8117 8.45E-05 0.04390 HSPA1A 1.5433 3.32E-09 1.68E-06 FOSB 1.5493 1.88E-05 0.00951 DUSP6 1.1689 5.27E-07 2.67E-04 CCL4 1.1057 8.66E-05 0.04390 EGR4 1.4966 8.59E-09 3.79E-06 RND3 0.8880 2.34E-05 0.01079 NR4A1 1.6955 7.58E-07 3.49E-04 FKBP11 -1.0158 1.84E-04 0.08205 NR4A1 1.4166 8.97E-09 3.79E-06 RGS1 1.1871 2.88E-05 0.01216 EGR4 1.6425 9.33E-07 3.94E-04 GM9958 -0.9378 2.12E-04 0.08205 PTGS2 1.3968 1.51E-08 5.88E-06 GADD45G 1.0496 3.97E-05 0.01547 RGS2 1.3091 1.09E-06 4.23E-04 NR4A1 1.1258 2.21E-04 0.08205 BTG2 1.2922 2.79E-08 1.01E-05 DUSP6 0.8763 4.39E-05 0.01547 JUN 1.1095 1.17E-06 4.23E-04 CCRL2 0.9681 2.27E-04 0.08205 EGR2 1.3976 4.45E-08 1.51E-05 GALNT12 -1.1040 4.58E-05 0.01547 AKAP12 0.9770 1.51E-06 5.10E-04 CCN1 1.2480 2.56E-04 0.08651 Page 6 of 15 Ramiro et al. BMC Neuroscience (2023) 24:7 Page 7 of 15 Fig. 1 Venn diagram of differentially expressed genes in the ischemic brain (FDR < 0.25) in all studied groups. The 14 common differentially expressed genes are shown through a STRING network revealing the main interactions between molecules. The strength of the correlation between two genes (nodes) 20 is reflected by the thickness of the edge. Diabetic: diabetic young male mice; Female: young female mice; Male: young male mice; Old: aged male mice created an enrichment map that shows pathways whose control. So, each group of animals included 8 ischemic genes are significantly enriched in transcriptomics data- animals and 4 sham-control animals. sets. Remarkably, the acute inflammatory response was After MCAO induction, 4 proteins were found to the biological process that encompassed more DEG, fol- be differentially expressed between the IP and the CL lowed by hematopoietic or lymphoid organ development, brain regions in male mice (FDR < 0.25), 14 in female hemopoiesis, MAPK cascade and leukocyte differentia- mice and 799 in diabetic mice. However, when compar- tion. Moreover, the enrichment map of biological pro- ing both regions, there were no differentially expressed cesses revealed 478 genes that were significantly enriched (FDR < 0.25) proteins in old mice. Moreover, there in 214 biological processes (Fig. 2B). Of these 214 bio- were no proteins differentially expressed (FDR < 0.25) logical processes, 39 were common in all groups, 14 were between the IP and the CL in sham animals of any of only dysregulated in females, 22 in diabetics, 8 in olds, 5 the groups. The top 15 proteins differentially expressed in males and 48 were found only in the integrated gene after cerebral ischemia for each group are shown in list of all groups (Fig. 2C). Furthermore, the main molec- Table 2. ular functions of these DEG were related to phosphopro- tein phosphatase activity, chemokine receptor binding, receptor ligand activity and transcription regulator activ- Integrative analysis of proteomic results ity, being this last one the molecular function with more We aimed to identify common proteins dysregulated in DEG, especially in females (Fig. 3A–B). the ischemic brain after stroke in all groups, as they will be playing a key role in stroke pathophysiology. How- Proteomics results ever, we found no proteins differentially expressed shared We also analyzed protein changes occurring early after between all the groups. Interestingly, females and dia- cerebral ischemia in the brain using mass spectrometry. betic mice were the only groups sharing dysregulated All samples in each group of animals passed the quality proteins, with 10 common proteins. (See figure on next page.) Fig. 2 Pathway enrichment analysis revealing the main biological processes of differentially expressed genes in the brain 2 h after cerebral ischemia. A Top 20 enriched biological processes; bar plot showing the number of differentially expressed genes of each altered biological process, sorted by increasing p-value from left to right. Each color indicates the corresponding transcriptomic dataset, and the purple color (integromics) corresponds to a subset of enriched pathways with combined evidence that were only detected by integrating the data of all groups. B Enrichment map of biological processes altered acutely after ischemic stroke. Nodes in the network represent biological processes, and similar biological processes with many common genes are connected. Nodes are colored according to the supporting omics datasets. Only pathways with 2 or more nodes were represented. C Venn diagram of the enriched biological processes in all studied groups. Diabetic: diabetic young male mice; Female: young female mice; Male: young male mice; Old: aged male mice Ramiro et al. BMC Neuroscience (2023) 24:7 Page 8 of 15 Fig. 2 (See legend on previous page.) Ramiro et al. BMC Neuroscience (2023) 24:7 Page 9 of 15 Fig. 3 Pathway enrichment analysis revealing the main molecular functions of differentially expressed genes in the brain 2 h after cerebral ischemia. A Top 20 enriched molecular functions; bar plot showing the number of differentially expressed genes of each altered molecular function, sorted by increasing p-value from left to right. Each color indicates the corresponding transcriptomic dataset, and the purple color (integromics) corresponds to a subset of enriched pathways with combined evidence that were only detected by integrating the data of all groups. B Enrichment map of molecular functions altered after ischemic stroke. Nodes in the network represent molecular functions and similar molecular functions with many common genes are connected. Nodes are colored according to the supporting omics datasets. Diabetic: diabetic young male mice; Female: young female mice; Male: young male mice; Old: aged male mice Discussion promising neuroprotection results obtained in pre-clin- The present study identifies key transcriptomic and pro- ical studies into the clinical setting [10, 43]. Taking all teomic changes triggered acutely after cerebral ischemia this into account, we aimed at evaluating similarities and in brain samples of mice of different groups: (1) young differences in protein and gene changes triggered acutely male mice, (2) young female mice, (3) aged male mice after cerebral ischemia in mice of four different groups and (4) diabetic young male mice, through mass spec- differing in sex, age and the presence of comorbidities to trometry and microarrays. Moreover, the transcriptomics elucidate if these differences could explain the off-target data generated has been integrated and analyzed together molecules of choice when using standard young male to reveal molecules and pathways with key roles in the models. hyperacute phase of stroke pathophysiology. Regarding the transcriptomics results, it is worth not- To date, various studies have devoted their efforts to ing that the number of DEG after cerebral ischemia is describe proteomic and transcriptomic changes that similar between males and females (61 and 77 respec- occur in the brain after stroke using mice models of cer- tively), while in old animals the number of DEG is ebral ischemia. These studies have provided valuable reduced to 24 and in diabetic animals there is an knowledge that has contributed to a better characteriza- important increase up to 699. In this line, our results tion of stroke pathology. This better comprehension of suggest that the post-ischemic expression level of tran- the disease has aided in selecting new candidates to be scripts might be attenuated with age. In fact, an age- studied as neuroprotectants and/or biomarkers [38–40]. dependent decline in transcriptional homeostasis has However, the vast majority of these studies are per- been described before and is being extensively studied, formed in young healthy male mice, with relatively fewer although to date is poorly understood [44, 45]. On the studies using females, and anecdotic studies using mice contrary, the expression level of transcripts after cer- with comorbidities and aged mice [10]. However, it has ebral ischemia could be exacerbated by the presence of been demonstrated that the response triggered after a comorbidities such as diabetes, indicating that cerebral stroke is affected by sex, age and comorbidities [41, 42]. ischemia might be altering a higher number of molecular Importantly, it has been suggested that this lack of pre- processes and pathways in that latter group of animals. cision in representing the population mostly affected by Considering the top 15 genes that are differentially stroke in pre-clinical studies, might be one of the rea- expressed in each group (regarding their FDR), most of sons explaining the systematic failure in translating the them are common in all four groups and mainly encode Ramiro et al. BMC Neuroscience Table 2 Top 15 proteins differentially expressed (FDR < 0.25) between the ipsilateral hemisphere and the contralateral hemisphere 2 h after cerebral ischemia induction FDR false discovery rate; logFC Logarithmic fold change (2023) 24:7 Young male mice, males Young female mice, females Diabetic young male mice, diabetic Aged male mice, old UNIPROT logFC P.Value FDR UNIPROT logFC P.Value FDR UNIPROT logFC P.Value FDR UNIPROT logFC P.Value FDR Q8CI32/Bag5 − 0.3250 2.82E-08 7.01E-05 Q8BKX1/Baiap2 0.4799 1.62E-05 0.0413 Q8VBY2/Camkk1 − 0.4021 1.55E-07 3.85E-04 P11798/Camk2a − 0.2119 9.70E-05 0.1205 Q8CGY8/Ogt − 0.2743 1.29E-04 0.1068 O70433/Fhl2 − 0.6287 5.24E-07 6.51E-04 Q6PHZ2/Camk2d − 0.1114 3.09E-04 0.2427 Q920P5/Ak5 − 0.2659 1.43E-04 0.1068 Q62318/Trim28 − 0.2786 1.21E-06 0.0010 O35927/Ctnnd2 − 0.0972 3.91E-04 0.2427 F6SEU4/Syngap1 0.4880 1.93E-04 0.1068 Q8R570/Snap47 − 0.4342 6.25E-06 0.0037 P68404/Prkcb − 0.2485 2.09E-04 0.1068 P27005/S100a8 0.7479 7.42E-06 0.0037 P08414/Camk4 − 0.3514 4.61E-04 0.1680 P08414/Camk4 − 0.5602 9.31E-06 0.0039 Q9Z2Y3/Homer1 0.2486 5.01E-04 0.1680 Q64337/Sqtm1 − 0.4070 1.33E-05 0.0041 E9Q0K9/Elmod1 − 0.3820 5.27E-04 0.1680 Q9WV34/Mpp2 − 0.2949 1.34E-05 0.0041 Q8BFT9/Svop 0.2494 7.22E-04 0.2046 Q8BFR5/Tufm − 0.3072 1.68E-05 0.0046 Q61329/Zfhx3 0.1805 8.62E-04 0.2200 Q8K0E8/Fgb 0.8275 1.99E-05 0.0049 Q8VEA4/Chchd4 0.2649 1.01E-03 0.2201 Q3UHD6/Snx27 − 0.5525 2.30E-05 0.0049 E9QK62/Ngef − 0.2693 1.04E-03 0.2201 Q68FF6/Git1 − 0.4539 2.38E-05 0.0049 P97822/Anp32e − 0.1824 1.19E-03 0.2332 Q920P5/Ak5 − 0.5258 2.73E-05 0.0052 O88737/Bsn 0.2131 1.28E-03 0.2332 P68404/Prkcb − 0.2960 3.87E-05 0.0069 P04627/Araf − 0.4653 4.54E-05 0.0070 Page 10 of 15 Ramiro et al. BMC Neuroscience (2023) 24:7 Page 11 of 15 for transcription factors, chemokines and GTPases, found clear differences between aged mice and the other which are known to be key players during the hyperacute groups regarding cell adhesion processes and regulation phase of cerebral ischemia [46–49]. For example, NPAS4 of protein kinase activity, which seem to be unaltered in encodes for a transcription factor whose expression aged animals. Protein kinases are involved in the majority is induced in various brain insults, including cerebral of cellular pathways, especially cell signaling and signal ischemia [50, 51]. NPAS4 indeed has a neuroprotective transduction [64, 65]. In addition, cell adhesion processes role in cerebral ischemia, probably being involved in cell are triggered after cerebral ischemia to mediate the infil- death and inflammatory response modulation [52]. In tration of leukocytes into the brain parenchyma [66]. addition, CCL3 is an active mediator in cerebral post- The fact that we did not find significant DEG involved in ischemic inflammation, upregulated in the ischemic core these two processes in aged mice, highlights the impor- [53], crucial for the induction of monocyte and neutro- tance of further studying their relationship with aging phil recruitment at inflammatory sites [54]. Moreover, and stroke response. In this regard, considering that aged upregulation of ATF3 transcript in the ischemic core dur- patients have poorer outcomes and higher rates of mor- ing the hyper-acute phase has previously been described tality after stroke, pharmacologically modulating mol- in young male mice [55]. In fact, this study revealed that ecules involved in these biological processes during the ATF3 upregulation might be related to mechanisms hyperacute phase of the disease in the aged population underlying mitigation of the neurotoxicity at an early might be an interesting approach to be evaluated in the stage after ischemia. In the same line, it has been shown future as neuroprotective strategies [67]. that NR4A1 mRNA expression is rapidly induced in the The presence of comorbidities such as obesity, diabe- ischemic core in a rat model of focal cerebral ischemia tes and hypertension are known risk factors for ischemic [56, 57]. Besides, it is thought that NR4A1 plays an stroke [16]. Moreover, patients with these comorbidities important role in the regulation of neuroinflammation have poorer outcomes after cerebral ischemia [68–70]. and in neural apoptosis after ischemic stroke [58, 59]. We found that diabetic mice have a significant increase From all the DEG, it is worth noting that 35 of them in the number of DEG after stroke, being 10 times higher are shared between males and females, but 27 are exclu- than in the other groups, which might suggest that cer- sive for females and 16 for males, highlighting that ebral ischemia triggers the activation and modulation divergences due to sex are an important factor to be con- of an augmented number of biological processes in that sidered. In fact, the existence of significant sexual dimor- group of animals. It has been reported before that after phism in stroke response has been described before. cerebral ischemia diabetic mice had pronounced sys- For example, genetic and epigenetic factors, differen- temic and vascular inflammation, augmented blood- tial activation of cell-death programs, cell-cell signaling brain barrier disruption, increased pro-inflammatory pathways, and systemic immune responses have been response, metabolic dysregulation, severe brain damage proposed as contributors to sex differences in ischemic and worse neurological deficits [71–73]. Our results sup- stroke [41]. Actually, various cell signaling pathways such port this idea, given that the number of DEG in diabetic as ischemia-induced cell death, mitochondrial metabo- mice related to these biological processes is significantly lism and inflammation are different between males and higher than in the other groups. Interestingly, we found females [60–62]. an increased number of DEG related to cell adhesion and In reference to old mice, our results show that there leukocyte activation that were exclusively dysregulated in is a significant reduction in the number of DEG after diabetic mice. In fact, it has been described that diabetic the insult, being most of the DEG shared with the other mice had an augmented number of infiltrating leukocytes groups and only 4 of them exclusively dysregulated in old in the brain after stroke, indicating that these cells might mice. Interestingly, a previous study exploring the effect be responsible for the exacerbated ischemic brain injury of age on the transcriptional response to ischemic stroke, in this group of animals [73]. reported that aged mice differentially regulated more The enrichment pathways analysis revealed that the genes than young mice 3 days after the event [63]. In inflammatory response and hemopoiesis were key bio- the present study, we have found that within the hypera- logical processes involved in the animal differences. The cute phase of the disease (2 h after the occlusion) aged inflammatory response triggered after the ischemic event animals have less DEG than their young counterparts, plays an important role in the progression of stroke. which might suggest that the transcriptional response It has been described that it has a dual role, as it exerts to cerebral ischemia might be a slower process in aged deleterious effects but also contributes to brain protec- mice, although at later time points the response might tion and repair of the damaged area [74–76]. Our results be exacerbated. Regarding the biological processes and show that males had a greater number of DEG related molecular functions altered after ischemic stroke, we to this process than females. It has been shown that sex Ramiro et al. BMC Neuroscience (2023) 24:7 Page 12 of 15 hormones can regulate the immune system, as the major- in comparison to all other groups, while in old animals ity of immune cells have receptors for these molecules we did not find any dysregulated protein, maybe due to [77]. In the same line, progesterone has been associated the early time-point in which samples were obtained. with a reduction of pro-inflammatory cytokines, while Regarding males and females, we found a similar number depending on the context estrogen can either promote of altered proteins in both groups, although it was higher or inhibit the inflammatory response [78–80]. Moreo- in females, as observed in the transcriptomics results. In ver, diabetes and aging exacerbate inflammatory mecha- general, our results suggest that in the studied time-point nisms, increasing the susceptibility of the blood-brain genes had a higher contribution to the observed altera- barrier to damage, consequently augmenting the risk of tions triggered after cerebral ischemia than proteins. One hemorrhagic transformation, reperfusion injury, and cer- possible reason for this might be the early time point ebral edema following stroke, and also countering endog- (2 h after occlusion) in which we evaluated the changes enous resilience and recovery mechanisms activated after triggered after ischemia. In this regard, performing a stroke [81, 82]. In this regard, there are various neuropro- similar study and analysis at later time points will give tective agents targeting inflammation that have shown complementary and valuable information to better char- promising results in pre-clinical studies but have failed acterize divergences and similarities in cerebral ischemia in clinical translation, such as Anakinra, Enlimomab and response between the studied groups. In addition to Natalizumab [83]. Considering our results, it is plausible this, various biological reasons could be influencing the to think that the differences observed in the regulation of observed divergence between altered proteins and genes inflammatory response might be one of the possible rea- such as differential post-transcriptional mechanisms, sons influencing the failure of translating the preliminary the tight regulation of translation, and differences in the results obtained into the whole population. turnover of proteins and mRNAs [90–92]. Given the Regarding hemopoiesis, various studies have revealed scarce amount of dysregulated proteins that we obtained that acute cardiovascular events like myocardial infarc- in 3 of the studied groups, we were not able to perform tion or stroke greatly impact hemopoiesis via sympathetic an integrative analysis. activation, glucocorticoid release, and the production Altogether, we show robust data supporting differences of various proinflammatory molecules [84–86]. Indeed, in the molecular responses triggered by cerebral ischemia the systemic number of innate immune cells acutely depending on the sex, age and the presence of comor- increases after stroke. It has been demonstrated that bidities. These divergences might be one of the reasons bone marrow hematopoietic stem cells (HSC) and down- explaining the choice of inappropriate therapeutic targets stream hemopoietic progenitors augment their activity and the posterior failure in clinical translation of the main after MCAO in mice, leading to an increased output of neuroprotective agents tested in pre-clinical studies up to inflammatory monocytes and neutrophils [87]. However, the date. To improve future study designs to make clini- it has been described that common cardiovascular risk cal implementation more effective, the Stroke Therapy factors such as arterial hypertension, hyperlipoproteine- Academic Industry Roundtable (STAIR) published some mia, and diabetes mellitus considerably alter hemopoietic recommendations to enhance the quality of preclini- processes [86]. In the same line, it has been demon- cal research. Remarkably, these recommendations give strated that age also influences hematopoiesis, given that special attention to the experimental model, species and stem cells become more poorly functional and there is a strain selection and emphasize the necessity of consider- reduction in T-cell production, which may contribute to ing both sexes when designing experiments. In addition, a decrease in immunity seen in the elderly [88]. Regard- they highly advocate including animals with comorbidi- ing sex differences, various studies revealed that sex hor- ties such as diabetes, obesity or hypertension as well as mones exert significant effects on hemopoiesis through aged animals in pre-clinical studies, to better repre- selectively expressed receptors in various populations of sent the human population that most likely will suffer a stem/progenitor cells, controlling their self-renewal, dif- stroke [17–19]. The results obtained in the present study ferentiation, and proliferation [89]. Thus, considering that robustly support the necessity of implementing these the inflammatory response and the hemopoiesis are key recommendations when designing pre-clinical studies processes triggered after stroke, the differences described to test neuroprotective agents. In this regard, molecules in these mechanisms depending on phenotypes might be to be targeted in pre-clinical studies should be selected ultimately influencing the stroke response, explaining the considering the differences and similarities that exist differences observed in the present study. between groups. Moreover, the presence of molecules The proteomics results are in agreement with the ones exclusively dysregulated in one group of animals might found in transcriptomics. Diabetic animals had an enor- open the door to future studies directed towards per- mous number of proteins altered in the ischemic core sonalized medicine in subcohorts of stroke patients. For Ramiro et al. BMC Neuroscience (2023) 24:7 Page 13 of 15 example, if the functional analysis of molecules that are Acknowledgements dysregulated in diabetic, males and females but not in old Not applicable. mice showed neuroprotective effects, a pharmacologic Author contributions activation of these molecules in aged ischemic brains LR, AS and JM conceived and designed the experiments. LR performed all might improve the functional outcome in elderly stroke study and supplied the Progenesis LC–MS® software, and LM, PG-R conducted animal experiments, supported by FM and AS, FC helped in the design of the patients. Conversely, if one of the many molecules dys- the proteomics experiments. JF, PG-R and LR analyzed transcriptomics and regulated in diabetic mice shows negative neurodegen- proteomics data from the discovery phase, JF performed the integrative analy‑ erative or even neurotoxic functions, a pharmacological sis. AR and JM supervised all the experiments and LR drafted the manuscript. inhibition in diabetic ischemic brains could bear neuro- All authors have critically reviewed the article content and approved it in its final version. All authors read and approved the final manuscript. protective effects. This approach will definitely provide a personalized management of stroke patients, by treating Funding each patient considering its clinical characteristics. This work has been funded by Instituto de Salud Carlos III (PI18/00804) and by the European Regional Development Fund (FEDER). Neurovascular Research This study stands for some limitations that had to be Laboratory takes part in the Spanish stroke research network RICORS-ICTUS. considered. First, regarding the experimental design, L.R is supported by a pre-doctoral fellowship from the Instituto de Salud ideally, young male mice should have been performed Carlos III (IFI17/00012). simultaneously with the other groups, to reduce inter- Availability of data and materials assay variability. In this line, to minimize the variability, The mass spectrometry proteomics data have been deposited to the each group of animals was analyzed separately and cor- ProteomeXchange Consortium via the PRIDE [93, 94] partner repository with the dataset identifier PXD032141, and the transcriptomics data in the Gene rected by its own control samples. Second, in the pre- Expression Omnibus (GEO) database with the dataset identifier GSE196266. sent study, we have incorporated 4 different groups of animals differing in sex, age or the presence of diabetes. Declarations However, it would have been also interesting to include additional groups such as aged females, diabetic females Human and animal ethics All animal procedures were performed in compliance with the Spanish legisla‑ or mice with other comorbidities such as hypertension to tion and in accordance with the Directives of the European Union and were enrich the obtained information. Third, time extrapola- approved by the Ethics Committee of the Vall d’Hebron Research Institute tions between mice and humans are poorly understood (protocol number 03/19 and 73/20). All experiments were conducted in a randomized manner and in adherence to the ARRIVE guidelines. and not well established. A better comprehension of this equivalence is needed to better comprehend the implica- Ethical approval and consent to participate tions of the results presented here. Moreover, the early Not applicable. time-point in which we have evaluated the changes trig- Consent for publication gered after ischemic stroke seems not to be the optimum Not applicable. to find proteomic divergences between groups. Finally, Competing interests infarct volume was not assessed in the present study due Authors declare no competing interests. to the lack of sensitivity of TTC (triphenyltetrazolium chloride) staining to determine the lesion volume 2 h Received: 6 September 2022 Accepted: 10 January 2023 after occlusion. All in all, in the present study we have explored simi- larities and differences in protein and gene changes in the brain after cerebral ischemia depending on age, sex References and the presence of diabetes. In this regard, we have 1. Benjamin EJ, Blaha MJ, Chiuve SE, Cushman M, Das SR, Deo R, et al. Heart revealed that the molecular response to stroke varies disease and stroke statistics-2017 update: a report from the American depending on the phenotype of the animals, reinforcing heart association. Circulation. 2017;135:e146-603. 2. Fiehler J, Gerloff C. Mechanical thrombectomy in stroke. Dtsch Arztebl Int. the necessity of following STAIR recommendations when 2015;112:830–6. designing future studies to improve “bench to bedside” 3. Hacke W, Kaste M, Bluhmki E, Brozman M, Dávalos A, Guidetti D, et al. translation. Thrombolysis with Alteplase 3 to 4.5 hours after acute ischemic stroke. New England J Med. 2008;359:1317–29. 4. Lo EH, Dalkara T, Moskowitz MA. Neurological diseases: mechanisms, chal‑ Supplementary Information lenges and opportunities in stroke. Nat Rev Neurosci. 2003;4:399–414. The online version contains supplementary material available at https://​doi.​ 5. Urra X, Abilleira S, Dorado L, Ribó M, Cardona P, Millán M, et al. Mechanical org/​10.​1186/​s12868-​023-​00775-7. thrombectomy in and outside the REVASCAT trial: insights from a concur‑ rent population-based stroke registry. Stroke. 2015;46:3437–42. 6. Fisher M. Characterizing the target of acute stroke therapy. Stroke. 1997. Additional file 1: Table S1. Top 30 differentially expressed genes https://​doi.​org/​10.​1161/​01.​STR.​28.4.​866. (FDR<0.25) between the infarcted hemisphere and the contralateral 7. Goenka L, Satyanarayana CRU, Suresh SK, George M. Neuroprotective healthy hemisphere 2 hours after cerebral ischemia induction exclusive agents in acute ischemic stroke—a reality check. Biomed Pharmaco‑ from each group of animals. therapy. 2019. https://​doi.​org/​10.​1016/j.​biopha.​2018.​11.​041. Ramiro et al. BMC Neuroscience (2023) 24:7 Page 14 of 15 8. O’Collins VE, Macleod MR, Donnan GA, Horky LL, Van Der Worp BH, How‑ 31. Szklarczyk D, Gable AL, Nastou KC, Lyon D, Kirsch R, Pyysalo S, et al. The ells DW. 1026 Experimental treatments in acute stroke. Ann Neurol. 2006. STRING database in 2021: customizable protein-protein networks, and https://​doi.​org/​10.​1002/​ana.​20741. functional characterization of user-uploaded gene/measurement sets. 9. Moretti A, Ferrari F, Villa RF. Neuroprotection for ischaemic stroke: current Nucleic Acids Res. 2021. https://​doi.​org/​10.​1093/​nar/​gkaa1​074. status and challenges. Pharmacol Ther. 2015;146:23–34. 32. Paczkowska M, Barenboim J, Sintupisut N, Fox NS, Zhu H, Abd-Rabbo D, 10. Sutherland BA, Minnerup J, Balami JS, Arba F, Buchan AM, Kleinschnitz et al. Integrative pathway enrichment analysis of multivariate omics data. C. Neuroprotection for ischaemic stroke: translation from the bench to Nat Commun. 2020. https://​doi.​org/​10.​1038/​s41467-​019-​13983-9. the bedside. Int J Stroke. 2012. https://​doi.​org/​10.​1111/j.​1747-​4949.​2012.​ 33. Brown MB. 400: a method for combining non-independent one-sided 00770.x. tests of significance. Biometrics. 1975. https://​doi.​org/​10.​2307/​25298​26. 11. Coco DL, Lopez G, Corrao S. Cognitive impairment and stroke in elderly 34. Carbon S, Douglass E, Good BM, Unni DR, Harris NL, Mungall CJ, et al. patients. Vasc Health Risk Manag. 2016;12:105–16. The gene ontology resource: enriching a gold mine. Nucleic Acids Res. 12. Smajlović D. Strokes in young adults: epidemiology and prevention. Vasc 2021;49:D325–34. Health Risk Manag. 2015. https://​doi.​org/​10.​2147/​VHRM.​S53203. 35. The Gene Ontology Consortium. Gene Ontology: tool for the unification 13. Virani SS, Alonso A, Benjamin EJ, Bittencourt MS, Callaway CW, Carson AP, of biology. Nat Gen. 2000;25:25–9. et al. Heart disease and stroke statistics—2020 update: a report from the 36. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. American heart association. Circulation. 2020. https://​doi.​org/​10.​1161/​ Cytoscape: a software environment for integrated models of biomolecu‑ CIR.​00000​00000​000757. lar interaction networks. Genome Res. 2003. https://​doi.​org/​10.​1101/​gr.​ 14. Appelros P, Stegmayr B, Terent A. Sex differences in stroke epidemiology: 12393​03. a systematic review. Stroke. 2009. https://​doi.​org/​10.​1161/​STROK​EAHA.​ 37. Merico D, Isserlin R, Stueker O, Emili A, Bader GD. Enrichment map: a 108.​540781. network-based method for gene-set enrichment visualization and inter‑ 15. Nichols M, Townsend N, Luengo-Fernandez R. European Cardiovascular pretation. PLoS ONE. 2010. https://​doi.​org/​10.​1371/​journ​al.​pone.​00139​84. Disease Statistics 2012. European Heart Network, Brussels, European 38. Chen H, He Y, Chen S, Qi S, Shen J. Therapeutic targets of oxidative/nitro‑ Society of Cardiology, Sophia Antipolis. 2012. Cerebrovasc Dis. 2012. sative stress and neuroinflammation in ischemic stroke: applications for 16. Boehme AK, Esenwa C, Elkind MSV. Stroke risk factors, genetics, and pre‑ natural product efficacy with omics and systemic biology. Pharmacol Res. vention. Circ Res. 2017. https://​doi.​org/​10.​1161/​CIRCR​ESAHA.​116.​308398. 2020. https://​doi.​org/​10.​1016/j.​phrs.​2020.​104877. 17. Fisher M. Recommendations for standards regarding preclinical neuro‑ 39. Neuhaus AA, Couch Y, Hadley G, Buchan AM. Neuroprotection in stroke: protective and restorative drug development. Stroke. 1999. https://​doi.​ the importance of collaboration and reproducibility. Brain. 2017. https://​ org/​10.​1161/​01.​STR.​30.​12.​2752. doi.​org/​10.​1093/​brain/​awx126. 18. Fisher M, Hanley DF, Howard G, Jauch EC, Warach S. Recommendations 40. Wang Y, Guo W, Xie S, Liu Y, Xu D, Chen G, et al. Multi-omics analysis of from the STAIR V meeting on acute stroke trials, technology and out‑ brain tissue metabolome and proteome reveals the protective effect of comes. Stroke. 2007. https://​doi.​org/​10.​1161/​01.​STR.​00002​55951.​37434.​ gross saponins of Tribulus terrestris L. fruit against ischemic stroke in rat. J aa. Ethnopharmacol. 2021. https://​doi.​org/​10.​1016/j.​jep.​2021.​114280. 19. Albers GW, Goldstein LB, Hess DC, Wechsler LR, Furie KL, Gorelick PB, et al. 41. Spychala MS, Honarpisheh P, McCullough LD. Sex differences in neuroin‑ Stroke treatment academic industry roundtable (STAIR) recommenda‑ flammation and neuroprotection in ischemic stroke. J Neurosci Res. 2017. tions for maximizing the use of intravenous thrombolytics and expand‑ https://​doi.​org/​10.​1002/​jnr.​23962. ing treatment options with intra-arterial and neuroprotective therapies. 42. Howells DW, Porritt MJ, Rewell SSJ, O’Collins V, Sena ES, Van Der Worp HB, Stroke. 2011;42:2645–50. et al. Different strokes for different folks: The rich diversity of animal mod‑ 20. Simats A, Ramiro L, García-Berrocoso T, Briansó F, Gonzalo R, Martín L, els of focal cerebral ischemia. J Cereb Blood Flow Metab. 2010. https://​ et al. A mouse brain-based multi-omics integrative approach reveals doi.​org/​10.​1038/​jcbfm.​2010.​66. potential blood biomarkers for ischemic stroke. Mol Cell Proteomics. 43. Xu SY, Pan SY. The failure of animal models of neuroprotection in acute 2020. https://​doi.​org/​10.​1074/​mcp.​RA120.​002283. ischemic stroke to translate to clinical efficacy. Med Sci Monitor Basic Res. 21. Kilkenny C, Browne WJ, Cuthill IC, Emerson M, Altman DG. Improving 2013. https://​doi.​org/​10.​12659/​msmbr.​883750. bioscience research reporting: the arrive guidelines for reporting animal 44. Gupta K, Yadav P, Maryam S, Ahuja G, Sengupta D. Quantification of age- research. PLoS Biol. 2013;4:35–44. related decline in transcriptional homeostasis. J Mol Biol. 2021. https://​ 22. Catherine Hagan. When are mice considered old?. 2017. https://​www.​jax.​ doi.​org/​10.​1016/j.​jmb.​2021.​167179. org/​news-​and-​insig​hts/​jax-​blog/​2017/​novem​ber/​when-​are-​mice-​consi​ 45. Perez-Gomez A, Buxbaum JN, Petrascheck M. The aging transcriptome: dered-​old. Accessed 11 Dec 2022. read between the lines. Curr Opin Neurobiol. 2020. https://​doi.​org/​10.​ 23. Clark W, Lessov N, Dixon M, Eckenstein F. Monofilament intraluminal mid‑ 1016/j.​conb.​2020.​05.​001. dle cerebral artery occlusion in the mouse. Neurol Res. 1997;19:641–8. 46. Lakhan SE, Kirchgessner A, Hofer M. Inflammatory mechanisms in 24. Liu S, Zhen G, Meloni BP, Campbell K, Winn HR. 2009 Rodent stroke model ischemic stroke: therapeutic approaches. J Transl Med. 2009;7:97–107. guidelines for preclinical stroke trials (1st edition). J Exp Stroke Transl Med. 47. Dirnagl U, Klehmet J, Braun JS, Harms H, Meisel C, Ziemssen T, et al. 2:2–27. Stroke-induced immunodepression: experimental evidence and clinical 25. Lee S, Lee M, Hong Y, Won J, Lee Y, Kang SG, et al. Middle cerebral artery relevance. Stroke. 2007;38(2):770–3. occlusion methods in rat versus mouse models of transient focal cerebral 48. Endres M, Dirnagl U, Moskowitz MA. Chapter 2 The ischemic cascade and ischemic stroke. Neural Regen Res. 2014. https://​doi.​org/​10.​4103/​1673-​ mediators of ischemic injury handbook of clinical neurology. Amsterdam: 5374.​131582. Elsevier; 2008. 26. Kauffmann A, Gentleman R, Huber W. arrayQualityMetrics—a bioconduc‑ 49. Shi GX, Andres DA, Cai W. Ras Family Small GTPase-Mediated Neuro‑ tor package for quality assessment of microarray data. Bioinformatics. protective Signaling in Stroke. Cent Nerv Syst Agents Med Chem. 2012. 2009;25:415–6. https://​doi.​org/​10.​2174/​18715​24117​96011​349. 27. Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, 50. Leong WK, Klaric TS, Lin Y, Lewis MD, Koblar SA. Upregulation of the neu‑ et al. Exploration, normalization, and summaries of high density oligonu‑ ronal Per-Arnt-Sim domain protein 4 (Npas4) in the rat corticolimbic sys‑ cleotide array probe level data. Biostatistics. 2003;4:249–64. tem following focal cerebral ischemia. Eur J Neurosci. 2013;37:1875–84. 28. Smyth GK. Linear models and empirical bayes methods for assessing dif‑ 51. Shamloo M, Soriano L, von Schack D, Rickhag M, Chin DJ, Gonzalez- ferential expression in microarray experiments. Stat Appl Genet Mol Biol. Zulueta M, et al. Npas4, a novel helix-loop-helix PAS domain pro‑ 2004;3:1–25. tein, is regulated in response to cerebral ischemia. Eur J Neurosci. 29. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical 2006;24:2705–20. and powerful approach to multiple testing. J Roy Stat Soc Ser B Meth‑ 52. Choy FC, Klarić TS, Koblar SA, Lewis MD. The role of the neuroprotective odol. 1995;57:289–300. factor Npas4 in cerebral ischemia. Int J Mol Sci. 2015;16:29011–28. 30. van den Berg RA, Hoefsloot HCJ, Westerhuis JA, Smilde AK, van der Werf 53. Cowell RM, Xu H, Galasso JM, Silverstein FS. Hypoxic-ischemic injury MJ. Centering, scaling, and transformations: Improving the biological induces macrophage inflammatory protein-1alpha expression in imma‑ information content of metabolomics data. BMC Genomics. 2006;7:1–15. ture rat brain. Stroke. 2002;33:795–801. Ramiro et al. BMC Neuroscience (2023) 24:7 Page 15 of 15 54. Mirabelli-Badenier M, Braunersreuther V, Viviani GL, Dallegri F, Quercioli 75. Ramiro L, Simats A, García-Berrocoso T, Montaner J. Inflammatory mol‑ A, Veneselli E, et al. CC and CXC chemokines are pivotal mediators of ecules might become both biomarkers and therapeutic targets for stroke cerebral injury in ischaemic stroke. Thromb Haemost. 2011;105:409–20. management. Ther Adv Neurol Disord. 2018;11:1–24. 55. Takarada T, Kou M, Hida M, Fukumori R, Nakamura S, Kutsukake T, et al. 76. Vidale S, Consoli A, Arnaboldi M, Consoli D. Postischemic inflammation in Protective upregulation of activating transcription factor-3 against acute stroke. J Clin Neurol. 2017;13:1–9. glutamate neurotoxicity in neuronal cells under ischemia. J Neurosci Res. 77. Buskiewicz IA, Huber SA, Fairweather DL. Sex hormone receptor expres‑ 2016;94:378–88. sion in the immune system. Sex Differences Physiol. 2016. https://​doi.​org/​ 56. Johansson IM, Wester P, Háková M, Gu W, Seckl JR, Olsson T. Early and 10.​1016/​B978-0-​12-​802388-​4.​00004-5. delayed induction of immediate early gene expression in a novel focal 78. Robertson CL, Fidan E, Stanley RM, Noje C, Bayir H. Progesterone for cerebral ischemia model in the rat. Eur J Neurosci. 2000;12:3615–25. neuroprotection in pediatric traumatic brain injury. Pediatr Crit Care Med. 57. Dahlqvist P, Rönnbäck A, Risedal A, Nergårdh R, Johansson I-M, Seckl JR, 2015. https://​doi.​org/​10.​1097/​PCC.​00000​00000​000323. et al. Effects of postischemic environment on transcription factor and 79. Liu M, Kelley MH, Herson PS, Hurn PD. Neuroprotection of sex steroids. serotonin receptor expression after permanent focal cortical ischemia in Minerva Endocrinol. 2010;35(2):127–43. rats. Neuroscience. 2003;119:643–52. 80. Di Florio DN, Sin J, Coronado MJ, Atwal PS, Fairweather DL. Sex differ‑ 58. Zhang Y-J, Song J-R, Zhao M-J. NR4A1 regulates cerebral ischemia- ences in inflammation, redox biology, mitochondria and autoimmunity. induced brain injury by regulating neuroinflammation through interac‑ Redox Biol. 2020. https://​doi.​org/​10.​1016/j.​redox.​2020.​101482. tion with NF-κB/p65. Biochem Biophys Res Commun. 2019;518:59–65. 81. Krinock MJ, Singhal NS. Diabetes, stroke, and neuroresilience: looking 59. Xiao G, Sun T, Songming C, Cao Y. NR4A1 enhances neural survival fol‑ beyond hyperglycemia. Ann N Y Acad Sci. 2021;1495:78–98. lowing oxygen and glucose deprivation: an in vitro study. J Neurol Sci. 82. Venkat P, Chopp M, Chen J. Blood-brain barrier disruption, vascular 2013;330:78–84. impairment, and ischemia/reperfusion damage in diabetic stroke. J Am 60. Demarest TG, McCarthy MM. Sex differences in mitochondrial (dys) Heart Assoc. 2017;6:6. function: Implications for neuroprotection. J Bioenerg Biomembr. 2015. 83. Karsy M, Brock A, Guan J, Taussky P, Kalani MYS, Park MS. Neuroprotective https://​doi.​org/​10.​1007/​s10863-​014-​9583-7. strategies and the underlying molecular basis of cerebrovascular stroke. 61. Hagberg H, Wilson MA, Matsushita H, Zhu C, Lange M, Gustavsson M, Neurosurg Focus. 2017;42:E3. et al. PARP-1 gene disruption in mice preferentially protects males from 84. Swirski FK, Nahrendorf M. Leukocyte behavior in atherosclerosis, myocar‑ perinatal brain injury. J Neurochem. 2004. https://​doi.​org/​10.​1111/j.​1471-​ dial infarction, and heart failure. Science. 2013;339:161–6. 4159.​2004.​02547.x. 85. Courties G, Frodermann V, Honold L, Zheng Y, Herisson F, Schloss MJ, et al. 62. Liu F, Li Z, Li J, Siegel C, Yuan R, Mccullough LD. Sex differences in caspase Glucocorticoids regulate bone marrow B lymphopoiesis after stroke. Circ activation after stroke. Stroke. 2009. https://​doi.​org/​10.​1161/​STROK​EAHA.​ Res. 2019;124:1372–85. 108.​538686. 86. Poller WC, Nahrendorf M, Swirski FK. Hematopoiesis and cardiovascular 63. Androvic P, Kirdajova D, Tureckova J, Zucha D, Rohlova E, Abaffy P, et al. disease. Circ Res. 2020;126:1061–85. Decoding the transcriptional response to ischemic stroke in young and 87. Courties G, Herisson F, Sager HB, Heidt T, Ye Y, Wei Y, et al. Ischemic aged mouse brain. Cell Rep. 2020. https://​doi.​org/​10.​1016/j.​celrep.​2020.​ stroke activates hematopoietic bone marrow stem cells. Circ Res. 107777. 2015;116:407–17. 64. Ardito F, Giuliani M, Perrone D, Troiano G, Muzio LL. The crucial role of 88. Groarke EM, Young NS. Aging and hematopoiesis. Clin Geriatr Med. protein phosphorylation in cell signalingand its use as targeted therapy 2019;35:285–93. (Review). Int J Mol Med. 2017. https://​doi.​org/​10.​3892/​ijmm.​2017.​3036. 89. Cui X, Zhao X, Liang Y. Sex differences in normal and malignant hemat‑ 65. Cheng HC, Qi RZ, Paudel H, Zhu HJ. Regulation and function of protein opoiesis. Blood sci Baltimore Md. 2022;4:185–91. kinases and phosphatases. Enzyme Res. 2011. https://​doi.​org/​10.​4061/​ 90. Fanara P, Wong PYA, Husted KH, Liu S, Liu VM, Kohlstaedt LA, et al. 2011/​794089. Cerebrospinal fluid-based kinetic biomarkers of axonal transport in moni‑ 66. Mizuma A, Yenari MA. Anti-inflammatory targets for the treatment of toring neurodegeneration. J Clin Investig. 2012. https://​doi.​org/​10.​1172/​ reperfusion injury in stroke. Front Neurol. 2017. https://​doi.​org/​10.​3389/​ JCI64​575. fneur.​2017.​00467. 91. Lam MPY, Wang D, Lau E, Liem DA, Kim AK, Ng DCM, et al. Protein kinetic 67. Virani SS, Alonso A, Aparicio HJ, Benjamin EJ, Bittencourt MS, Callaway signatures of the remodeling heart following isoproterenol stimulation. J CW, et al. Heart disease and stroke statistics—2021 update. Circulation. Clin Investig. 2014. https://​doi.​org/​10.​1172/​JCI73​787. 2021. https://​doi.​org/​10.​1161/​cir.​00000​00000​000950. 92. Vogel C, Marcotte EM. Insights into the regulation of protein abun‑ 68. Schmidt M, Jacobsen JB, Johnsen SP, Boøtker HE, Soørensen HT. dance from proteomic and transcriptomic analyses. Nat Rev Genet. Eighteen-year trends in stroke mortality and the prognostic influence of 2012;13:227–32. comorbidity. Neurology. 2014. https://​doi.​org/​10.​1212/​WNL.​00000​00000​ 93. Deutsch EW, Bandeira N, Sharma V, Perez-Riverol Y, Carver JJ, 000062. Kundu DJ, et al. The ProteomeXchange consortium in 2020: enabling “big 69. Dubow J, Fink ME. Impact of hypertension on stroke. Curr Atheroscler data” approaches in proteomics. 2020. Nucleic Acids Res. https://​doi.​org/​ Rep. 2011. https://​doi.​org/​10.​1007/​s11883-​011-​0187-y. 10.​1093/​nar/​gkz984. 70. Hu G, Sarti C, Jousilahti P, Peltonen M, Qiao Q, Antikainen R, et al. The 94. Perez-Riverol Y, Bai J, Bandla C, Hewapathirana S, García-Seisdedos D, impact of history of hypertension and type 2 diabetes at baseline on the Kamatchinathan S, Kundu D, Prakash A, Frericks-Zipper A, Eisenacher M, incidence of stroke and stroke mortality. Stroke. 2005. https://​doi.​org/​10.​ Walzer M, Wang S, Brazma AVJ. The PRIDE database resources in 2022: a 1161/​01.​STR.​00001​90894.​30964.​75. hub for mass spectrometry-based proteomics evidences. Nucleic Acids 71. Ma S, Wang J, Wang Y, Dai X, Xu F, Gao X, et al. Diabetes mellitus impairs Res. 2022. https://​doi.​org/​10.​1093/​nar/​gkab1​038. white matter repair and long-term functional deficits after cerebral ischemia. Stroke. 2018. https://​doi.​org/​10.​1161/​STROK​EAHA.​118.​021452. 72. Jiang Y, Liu N, Wang Q, Yu Z, Lin L, Yuan J, et al. Endocrine regulator Publisher’s Note rFGF21 (recombinant human fibroblast growth factor 21) improves Springer Nature remains neutral with regard to jurisdictional claims in pub‑ neurological outcomes following focal ischemic stroke of type 2 diabetes lished maps and institutional affiliations. mellitus male mice. Stroke. 2018. https://​doi.​org/​10.​1161/​STROK​EAHA.​ 118.​022119. 73. Zhang F, Zhao Q, Jiang Y, Liu N, Liu Q, Shi FD, et al. Augmented brain infiltration and activation of leukocytes after cerebral ischemia in type 2 diabetic mice. Front Immunol. 2019. https://​doi.​org/​10.​3389/​fimmu.​2019.​ 02392. 74. Simats A, García-Berrocoso T, Montaner J. Neuroinflammatory biomarkers: from stroke diagnosis and prognosis to therapy. Biochim Biophys Acta Mol Basis Dis. 2016;1862:411–24.