Liver proteome alterations in psychologically distressed rats and a nootropic drug

Background Chronic psychological distress is considered today a pandemic due to the modern lifestyle and has been associated with various neurodegenerative, autoimmune, or systemic inflammation-related diseases. Stress is closely related to liver disease exacerbation through the high activity of the endocrine and autonomic nervous systems, and the connection between the development of these pathologies and the physiological effects induced by oxidative stress is not yet completely understood. The use of nootropics, as the cognitive enhancer and antioxidant piracetam, is attractive to repair the oxidative damage. A proteomic approach provides the possibility to obtain an in-depth comprehension of the affected cellular processes and the possible consequences for the body. Therefore, we considered to describe the effect of distress and piracetam on the liver proteome. Methods We used a murine model of psychological stress by predatory odor as a distress paradigm. Female Sprague-Dawley rats were distributed into four experimental groups (n = 6 − 7/group) and were exposed or not to the stressor for five days and treated or not with piracetam (600 mg/kg) for six days. We evaluated the liver proteome by one-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis (1D-SDS-PAGE) followed by liquid chromatography-tandem mass spectrometry (GeLC-MS/MS). Besides, we analyzed the activity of liver antioxidant enzymes, the biochemical parameters in plasma and rat behavior. Results Our results showed that distress altered a wide range of proteins involved in amino acids metabolism, glucose, and fatty acid mobilization and degradation on the way to produce energy, protein folding, trafficking and degradation, redox metabolism, and its implications in the development of the non-alcoholic fatty liver disease (NAFLD). Piracetam reverted the changes in metabolism caused by distress exposure, and, under physiological conditions, it increased catabolism rate directed towards energy production. These results confirm the possible relationship between chronic psychological stress and the progression of NAFLD, as well as we newly evidenced the controversial beneficial effects of piracetam. Finally, we propose new distress biomarkers in the liver as the protein DJ-1 (PARK7), glutathione peroxidase 1 (GPX), peroxiredoxin-5 (PRDX5), glutaredoxin 5 (GLRX5), and thioredoxin reductase 1 (TXNDR1), and in plasma as biochemical parameters related to kidney function such as urea and blood urea nitrogen (BUN) levels.

288 according the manufacturer's procedure. For the protein identification, 50 µg of total proteins 289 extracted from each liver were pooled (n=6-7) for each experimental group (Fig. S1C). 290 291 GeLC-MS/MS analysis 292 The protein identification was performed using one-dimensional sodium dodecyl sulfate-293 polyacrylamide gel electrophoresis (1D-SDS-PAGE) coupled by liquid chromatography-tandem 294 mass spectrometry (GeLC-MS/MS) analysis (Fig. S1C). A total of 100 μg of pooled proteins 295 from each experimental group was pre-separated by 1D-SDS-PAGE, gels were stained by the 296 Coomassie method, and finally, each gel band was dissected manually in three equal parts. In-gel 297 protein digestion and nano LC-MS/MS were performed according to the methodology previously 298 described (Espinosa-Gómez et al., 2020) (for the complete description of the methodology 299 carried out in this analysis, see Methodology S1). Briefly, firstly, each gel section was distained, 300 and then dehydrated. Finally, proteins in gel sections were reduced with DTT and alkylated with 301 iodoacetamide. For in-gel digestion, the gels sections were incubated in a solution containing 302 12.5 ng/μL mass spectrometry grade Trypsin Gold (Promega, Madison, WI, USA) in 5 mM 303 NH 4 HCO 3 , at 37 °C overnight. Peptide samples were analyzed using an UltiMate 3000 304 RSnanoLC system (Thermo-Fisher Scientific, San Jose, CA) interfaced with Orbitrap Fusion TM 305 Tribid TM (Thermo-Fisher Scientific, San Jose, CA) mass spectrometer. All MS data were 306 obtained through Xcalibur 4.0.27.10 software (Thermo-Fisher Scientific) (Espinosa-Gómez et 307 al., 2020). Each sample of pooled proteins for each experimental group was run on one time. 308 309 Proteomics data analysis and biological interpretation 310 Mass spectra were analyzed with Proteome Discoverer 2.1 software (PD, Thermo Fisher 311 Scientific Inc.). The later searches were performed through Mascot server (version 2.4.1, Matrix 312 Science, Boston, MA), SEQUEST HT (Eng, McCormack & Yates, 1994), and AMANDA 313 (Dorfer et al., 2014). For protein identification, 25, 3.7, and 200 scores were considered, 314 respectively, for each search engine. The search with each engine was conducted against the 315 UniProt rat reference proteome (R. novergicus) database 316 (https://www.uniprot.org/proteomes/UP000002494) (for the complete description of the protein 317 search parameters in this analysis, see Methodology S1). The exponentially modified protein 318 abundance index (emPAI) was calculated for the label-free relative quantitation of the proteins in 319 each sample analyzed by nanoLC-MS/MS (Shinoda, Tomita & Ishihama, 2009). Only proteins 320 identified with at least two unique peptides and a 1.5-fold difference in their relative abundance 321 (0.66  fold change  1.50) were considered for further analysis. 322 Gene Ontology (GO) analysis and protein network were performed using the Search Tool for the 323 Retrieval of Interacting Genes/Proteins (STRING) (https://string-324 db.org/cgi/input?sessionId=b8yhKk9BbQSn&input_page_show_search=on) (Szklarczyk et al., 325 2019). Enrichment pathway analysis was carried out using the Gene Annotation & Analysis 326 Resource Metascape (http://metascape.org/gp/index.html#/main/step1) (Zhou et al., 2019). The 327 enrichment cluster analysis was made using the following settings: p-value < 0.01, minimum 328 count of 3, enrichment factor > 1.5, and FDR > 0.05. 329 330 Statistical analysis 331 The normality of the animal behavior, the antioxidant enzyme activity, and the biochemical 332 plasma data were tested using the Kolmogorov-Smirnoff (KS) test. Furthermore, for the 333 descriptive data, one-way ANOVA and Pearson's correlation coefficient (rho) were determined, 334 considering a statistically significant of p ≤ 0.05. The statistical analysis was performed using 335 SPSS v.8.0 software (SPSS Inc. Chicago, IL, USA). 336 337 Results 338 Protein identification by GeLC-MS/MS approach 339 To evaluate the effect of the chronic psychological distress and piracetam on female rat liver 340 proteome, proteins were extracted using a TCA-acetone-based method. Protein samples from 341 each liver were pooled (n=6-7) and pools from each experimental condition were subjected to a 342 pre-separation by 1D SDS-PAGE. Then, proteins were in-gel digested by trypsin, extracted from 343 gel pieces and analyzed by a shotgun proteomic approach. Each experimental group was 344 analyzed independently by label-free relative quantitation using the emPAI values. A total of 345 1,302 proteins were identified with at least one unique peptide and 1 % FDR (Data S1). Of these, 346 894 proteins with at least two unique peptides were considered for further analysis. The 347 intersection of proteins identified in all the experimental conditions was shown by a Venn 348 diagram (Fig. 1A). A 1.5-fold change threshold was used as cut-off values for protein abundance 349 changes (50 % of value variation). According to this criterion, a total of 350 proteins exhibited 350 qualitative differences between the four conditions: 112 proteins were found unique in one of the 351 four experimental conditions, 103 were presented only in two of the four groups, and 135 in 352 three groups (Fig. 1A). Regarding the quantitative differences, 313 proteins showed a change in 353 their abundance comparing S+P− versus S−P−, 259 proteins in S+P− versus S+P+, and 315 354 proteins in S−P+ versus S−P− (Data S2). 355 Gene Ontology through STRING search tool annotation was used to determine the subcellular 356 localization of all proteins identified. The most highly represented categories in this study were 357 membrane-associated, mitochondrial, cytosolic, and nuclear proteins, counting more than 100 358 proteins in each category (Fig. 1B). 359 360 Biological interpretation of differentially accumulated proteins 361 Proteins function was analyzed using Kyoto Encyclopedia of Genes and Genomes (KEGG) 362 annotation through Metascape online tool (Fig. 2). Liver from rats exposed to distress (S+P−) 363 showed significant enrichment in 16 pathways in which proteins were less abundant, and in 31 364 pathways in which proteins were more abundant compared to the liver from rats not exposed to 365 distress (S−P−) ( Fig. 2A). In general, proteins related to amino acids metabolism, glucose and 366 fatty acid (FA) mobilization and degradation to produce energy (glycolysis/gluconeogenesis, 367 TCA cycle, pyruvate metabolism, FA degradation, PPAR signaling pathway, dicarboxylate 368 metabolism, butanoate metabolism), protein folding, trafficking and degradation, metabolism of 369 xenobiotics by cytochrome P450, GSH metabolism, antioxidant defense mechanisms, and non-370 alcoholic fatty acid liver or Parkinson's diseases, among others, were highly affected in distress 371 (S+P−) when compared to S−P− ( Fig. 2A, Data S3).
372 When piracetam was supplied to the distressed group (S+P+), 24 and 19 pathways showed a 373 significant enrichment in which proteins were less and more abundant, respectively, compared to 374 S+P− (Fig. 2B). Piracetam reverted the changes in metabolism caused by distress exposure (Fig.  375 2A and 2B, Data S3). When piracetam was administrated to non-distressed rats (S−P+), 16 and 376 34 pathways showed a significant enrichment in which proteins were less and more abundant, 377 respectively, compared to the control group (S−P−) (Fig. 2C). The functional enrichment 378 analysis showed that proteins involved in amino acid metabolism, both biosynthesis and 379 especially degradation pathways, and in the oxidation of glucose and fatty acids were more 380 abundant. In the same way, proteins involved in the metabolism of xenobiotics by cytochrome 381 P450 and GSH showed an increase in their abundance ( Fig. 2A and 2C, Data S3). 382 This study highlights that psychological distress and piracetam induce changes in protein 383 abundances that are involved in redox metabolism, non-alcoholic fatty liver disease (NAFLD), 384 ER stress, and metabolism of xenobiotics by Cyt P450. 385 386 Altered proteins involved in redox metabolism 387 In this study, a total of 26 proteins involved in redox homeostasis were detected in the hepatic 388 proteome, of which 19 of them presented differences in 50 % in protein abundance changes 389 between S+P− versus S−P−, S+P+ versus S+P−, and S−P+ versus S−P− (0.66  fold change  390 1.50) ( Table S1). 391 To validate the changes in redox enzymes found in the proteomic analysis, the activity of the 392 selected antioxidant enzyme was also measured by biochemical methods (Table 1, Data S4). 393 Although no change in GPX protein abundance was shown in distressed rats (Table S1), a 394 statistically significant increase about two-fold in the GPX activity was observed in this 395 condition (p = 0.026) ( Table 1), and piracetam decreased this effect by 10 %, but without 396 resetting the control group values (Tables 1 and S1); however, the drug increased the protein 397 abundance in 75 % in the distressed rats (S+P+) ( Table S1). When piracetam was administered to 398 non-distressed animals (S−P+) also induces a statistically significant rise of more than 70 % in 399 GPX activity. 400 An increase of 50 % in CAT protein abundance was observed in rat liver exposed to stress 401 stimuli (Table S1); however, no significant changes in CAT activity were shown in this group 402 (Table 1). Piracetam induced a statistically significant decrease of 20 % in the CAT activity in 403 distressed group (S+P+), and this reduction was about 25 % in non-distressed animals (S−P+) (p 404 = 0.044) ( Table 1). The drug enhanced the CAT abundance by 18 % in the distressed group 405 (S+P+) compared to S+P− group and induced its increase for about 40 % in non-distressed 406 animals (S−P+). 407 Distress and piracetam induced no changes in total GST activity (Table 1); however, ten GST 408 isoforms were found in the proteomic analysis with different trend in protein abundance changes 409 (Table S1). Glutathione S-transferases, specifically glutathione S-transferase A4-4 (GSTA4), 410 keep in balance the cellular 4-hydroxynonenal (4-HNE) concentration, an end-product of n-6 411 PUFAs peroxidation related to increased apoptosis and necrosis, through its conjugation to GSH 492 ( Table 2, Data S4). The exposure to chronic distress induced a statistically significant decrease in 493 glucose levels in the distressed group (S+P−) (p = 0.012), and the administration of piracetam 494 avoided such an effect. The same happened in the triglycerides level, albeit not significant 495 changes. The specific biomarkers for hepatic injury and function were not affected by distress 496 and/or piracetam, although the total bilirubin presented a tendency to decrease in distressed 497 animals (S+P−) and with piracetam (S+P+) (p = 0.065). Moreover, a statistically significant 498 increase in the BUN level was found in distressed animals (S+P−), and piracetam counteracted 499 this effect reducing such levels, even more in rats exposed to distress (S+P−) (p = 0.025). Also, 500 piracetam significantly decreased plasma urea levels in distressed rats (S+P+) compared to 501 untreated rats (S+P−) (p = 0.025). 502 503 Correlational analysis among treatments, behavior, biochemical parameters, and 504 antioxidant enzyme activity 505 To study the association between treatments and behavior, a normality test of the variables was 506 first performed using Kolmogorov-Smirnoff (KS). Normal data distribution was shown and 507 therefore the statistical test Pearson's correlation coefficient (rho) was used, considering a 508 statistically significant correlation p ≤ 0.05. A high correlation was found between treatments 509 and behaviors ( Table S4). The time spent hiding was positively correlated with the treatments (r 510 = 0.82, p ≤ 0.00001), while the time spent in head-out position, exploring and approaching the 511 piece of cloth were negatively correlated with the treatments (r = -0.73, r = -0.83, and r = -0.72, 512 respectively; p ≤ 0.0001) (Fig. S2). These correlations showed that piracetam had no effect on 513 distressed rats but induced defensive behavior in non-distressed rats (Grigoruţă et al., 2018). 514 Further, correlational studies among biochemical parameters, antioxidant enzyme activity, 515 treatments, and rat behavior were performed. The indirect and total bilirubin were negatively 516 correlated with the treatments (r = -0.63, p = 0.02 and r = -0.58, p = 0.01, respectively) (Table  517 S4). Moreover, glucose level was negatively correlated with the time spent hiding (r = -0.49, p = 518 0.04) (Fig. S2A, Table S4), while it was positively correlated with the time spent in head out 519 position (r = 0.57, p = 0.04) (Fig. S2B, Table S4); direct bilirubin level was positively correlated 520 to the time spent approaching and exploring (r = 0.67, p = 0.01 and r = 0.59, p = 0.03, 521 respectively) (Fig. S2C, D, Table S4); and total bilirubin level was negatively correlated with the 522 time spent hiding (r = -0.45, p = 0.05) (Fig. S2E, Table S4), and positively correlated with the 523 time spent exploring (r = 0.45, p = 0.05) (Fig. S2F, Table S4). These correlations between 524 behavior and biochemical parameters show that bilirubin levels are related to stress behavior and 525 can be useful as a biomarker of psychological distress. 526 The correlational analysis among antioxidant enzyme activity with rat behavior and treatments 527 showed that only GPX activity had a positive association with the treatments (r = 0.61, p = 0.01) 528 (Table S4) and with the time spent hiding (r = 0.55, p = 0.01) (Fig. S2G), and it had negative 529 association with the exploration time (r = -0.61, p = 0.01) (Fig. S2H). These correlations support 530 the relationship of this enzyme with the stress behavior of hiding and it can be considered in 531 future studies as a distress biomarker. 532 533 Discussion 534 The connection between the physiological effects induced by oxidative stress and inflammation 535 caused by psychological distress and the associated pathologies is still unclear how occurs, as 536 well as the questionable beneficial effects of piracetam. The proteomic analysis provides a global 537 view of the affected cellular processes and the possible consequences on the body. Thus, a 538 GeLC-MS/MS analysis followed by a label-free relative quantitation based emPAI values of 539 identified proteins were carried out. Overall, our results showed that distress affected a wide 540 range of proteins involved in amino acid metabolism, glucose, and fatty acid (FA) mobilization 541 and degradation to produce energy, protein folding, trafficking and degradation, redox 542 metabolism, and NAFLD development. Piracetam reverted the changes in metabolism caused by 543 distress exposure, and, under physiological conditions, the drug induced an increased catabolism 544 rate directed towards energy production. 545 Physiologically, the body's primary response to stress is the fight or flight response that, at the 546 metabolic level, involves an increase of the energy production rate (increase ATP). After 547 exposure to a stressor, the hypothalamic-pituitary-adrenal axis is activated, causing the 548 production of the so-called stress hormones, such as epinephrine and cortisol (corticosterone in 549 murine), which in turn cause an increase of glycogenolysis, gluconeogenesis, lipid mobilization, 550 and inhibition of protein synthesis. Also, ATP synthesis leads to ROS generation, which can 551 promote oxidative stress, causing cellular damage to DNA, lipids, and proteins (Schiavone et al., 552 2013;Dumbell, Matveeva & Oster, 2016). Previous studies showed that emotional stress induces 553 oxidative damage and altered metabolism in the liver in different murine psychological stress 554 models (Depke et al., 2008(Depke et al., , 2009Jafari et al., 2014). 555 In the present study, we focused on liver proteins involved in redox metabolism, NAFLD, ER 556 stress, and metabolism of xenobiotics by Cyt P450, which were changed by psychological 557 distress and piracetam. 558 559 Redox metabolism 560 The maintenance of optimal ROS levels is essential in the body due to the dual function of these 561 species. ROS are signaling molecules that activate transcription factors that regulate the gene 562 expression related to growth and cell differentiation, but, on the other hand, ROS excess has been 563 associated with the development of numerous diseases such as type 2 diabetes, autoimmune 564 diseases, cancer or neurodegenerative diseases, among all (  representing the subcellular localization of identified proteins performed using STRING webtool. Most of the detected proteins in the four experimental groups were located within membrane, cytosol, mitochondria, and nucleus. Protein number is referred to as the count in a gene set, in this case, in the GO-term. S−P−, rats neither exposed to stress nor with piracetam treatment; S−P+, rats not exposed to stress but treated with piracetam; S+P−, rats exposed to stress without piracetam treatment; S+P+, rats exposed to stress and treated with piracetam.

Figure 2
Pathway enrichment analysis of altered proteins in the liver of rats exposed to distress and piracetam.  Interaction network of selected proteins with changes in their abundance in the liver of rats exposed to distress and piracetam.
Proteins involved in redox metabolism, GSH metabolism, PD, NAFLD, PPAR signaling pathway, protein processing in ER, metabolism of xenobiotics by Cyt P450, and apoptosis were grouped into sixteen clusters using STRING database. The protein interaction network showed a close relationship among these pathways. Clusters were grouped into five sets distributed around a central group with proteins involved in redox status homeostasis. Edges represent protein-protein associations made using STRING database with a medium confidence level (0.4). Protein clustering was carried out by the Markov Cluster Algorithm (MCL) with a specific inflation parameter of 3. The different colors in circles indicate diverse clusters. Solid lines represent protein interactions that belong to the same cluster. Dashed lines represent protein interactions belonging to a distinct cluster. The line thickness shows the strength of evidence, with thicker connections presenting higher confidence in the protein-protein interaction (the detailed description of the clusters is shown in Table S4). The colored ellipses show the regrouping of the closest clusters. Manuscript to be reviewed Delta body and liver weight and biochemical parameters in plasma obtained from each experimental group.
* Delta weight was calculated as the rat's body weight on the fifth day minus its body weight on the first day of the experiment and is shown in absolute value. † Liver value was calculated as the rat's liver weight (g) per 100 g of rat's body weight on the fifth day of the experiment. AST: aspartate aminotransferase; ALT: alanine aminotransferase; BUN: blood urea nitrogen; S−P−, rats neither exposed to stress nor with piracetam treatment; S−P+, rats not exposed to stress but treated with piracetam; S+P−, rats exposed to stress without piracetam treatment; S+P+, rats exposed to stress and treated with piracetam.