Alterations in the Proteome and Phosphoproteome Profiles of Rat Hippocampus after Six Months of Morphine Withdrawal: Comparison with the Forebrain Cortex

The knowledge about proteome changes proceeding during protracted opioid withdrawal is lacking. Therefore, the aim of this work was to analyze the spectrum of altered proteins in the rat hippocampus in comparison with the forebrain cortex after 6-month morphine withdrawal. We utilized 2D electrophoretic workflow (Pro-Q® Diamond staining and Colloidal Coomassie Blue staining) which was preceded by label-free quantification (MaxLFQ). The phosphoproteomic analysis revealed six significantly altered hippocampal (Calm1, Ywhaz, Tuba1b, Stip1, Pgk1, and Aldoa) and three cortical proteins (Tubb2a, Tuba1a, and Actb). The impact of 6-month morphine withdrawal on the changes in the proteomic profiles was higher in the hippocampus—14 proteins, only three proteins were detected in the forebrain cortex. Gene Ontology (GO) enrichment analysis of differentially expressed hippocampal proteins revealed the most enriched terms related to metabolic changes, cytoskeleton organization and response to oxidative stress. There is increasing evidence that energy metabolism plays an important role in opioid addiction. However, the way how morphine treatment and withdrawal alter energy metabolism is not fully understood. Our results indicate that the rat hippocampus is more susceptible to changes in proteome and phosphoproteome profiles induced by 6-month morphine withdrawal than is the forebrain cortex.


Introduction
Morphine is still considered a frequently used opioid in the treatment of moderate to severe pain. However, repetitive clinical use has many negative side effects [1,2]. The analgesic effect is caused by the activation of the opioid receptors (ORs). It has a high affinity for µ-OR (MOR) and a lower affinity for κ-OR (KOR) and δ-OR (DOR) [3,4].
During the last years, we published several animal studies related to the consequences of morphine treatment and withdrawal. Bourova et al. [5] described that exposure of rats to increasing doses of morphine (10-50 mg/kg, 10 days) results in significant desensitization of µ-ORand δ-OR-stimulated G protein response in the rat forebrain cortex. These findings were in agreement with the data published previously [6][7][8][9][10][11]. Ujcikova et al. [12] detected specific increased level of adenylyl cyclase I (8-fold) and adenylyl cyclase II (2.5-fold) in rat brain cortical plasma membrane samples after 10-day morphine treatment which returned to control level after 20 days of morphine withdrawal. There was no change in the expression level of other adenylyl cyclase isoforms (III-IX). Quantitative immunoblot analysis indicated the unchanged level of G protein α and β subunits: Gα i1 /Gα i2 , Gα i3 , Gα o , Gα q /Gα 11 , Gα s , Gα z , and Gβ. The same applied to Na, K-ATPase, and caveolin-1 [12,13].

Morphine Administration and Drug Withdrawal of Male Wistar Rats
Rats (8 weeks of age) were exposed to increasing doses of morphine (dissolved in 0.9% NaCl) for 10 days (10-50 mg/kg) in parallel with corresponding control animals according to our previously established protocols [5,12,[14][15][16][17]28] approved by the Ministry of Education, Youth and Sports of the Czech Republic (license number MSMT-1479/2019-6). Male Wistar rats were housed in the group of 3 per plastic cage on a 12/12 light/dark cycle. Food and water were available ad libitum. Described procedures were performed in an agreement with national and institutional guidelines for the care and use of animals in laboratory research.

Preparation of Samples
According to our experience and considering the sufficient amount of biological material needed for analyses, we used 9 animals of each testing group for the isolation of the brain cortex, hippocampus, striatum and cerebellum. The tissue from 3 randomly selected animals within the same group was pooled into one sample to obtain three equal amounts of brain tissues.

Detection of Phosphoproteins by Pro-Q ® Diamond Staining
Isoelectric focusing of samples containing 1 mg protein was performed according to our previously established scheme: 150 V for 5 h, 500 V for 1 h, 3500 V for 12 h, and 500 V for 3 h [15][16][17]. SDS-PAGE was followed by gel fixation in 250 mL of 50% methanol/10% acetic acid for 30 min and overnight with gentle agitation [30]. After three times washing in ultrapure water for 10 min, the gels were incubated in 160 mL of Pro-Q ® Diamond stain solution for 120 min in the dark. Pro-Q ® Diamond gel destaining solution was used three times for 30 min in the dark [17].

Staining by Colloidal Coomassie Blue (CBB)
CBB staining was used for detection of protein spots on 2D gels and subsequent mass spectrometric analysis as described in [14][15][16].

Statistical Analysis
The PDQuest TM software (Bio-Rad, version 7.3.1, Hercules, CA, USA) was used for evaluation of 2D gels. Protein spots were then checked manually. Relative abundances of protein spots showing significant quantitative differences at least 1.4-fold (p ≤ 0.05) were selected for mass spectrometric analysis. p-values were calculated by using unpaired Student's t-test and GraphPadPrism 8.3.0. Gene Ontology (GO) enrichment analysis of proteomic and phosphoproteomic profiles for rat brain hippocampus and cortex was performed using ShinyGO v0.74 tool in 20 October 2021 (bioinformatics.sdstate.edu/go, accessed on 20 October 2021); the p-value cut-off (FDR) was set to 0.05 for biological processes.
According to the current annotations (https://www.uniprot.org) in the UniProt database, the identified hippocampal proteins were involved in metabolism (5), cytoskeleton organization (4), signal transduction (

GO Enrichment Analysis of Altered Hippocampal Proteins
GO enrichment analysis of 20 significantly differentially expressed hippocampal proteins was carried out using the ShinyGO v0.74 tool (bioinformatics.sdstate.edu/go). The top thirty most significantly enriched GO terms for biological processes were summarized in hierarchical clustering tree, see Figure 3. The most enriched GO terms were related to metabolic changes: phosphorus metabolic process, phosphate-containing compound metabolic process, ATP metabolic process, methylglyoxal metabolic process, generation of precursor metabolites and energy, nucleoside phosphate metabolic process, nucleotide metabolic process; cytoskeleton organization: postsynaptic cytoskeleton organization, establishment of localization in cell, postsynaptic actin cytoskeleton organization, and oxidative stress: removal of superoxide radicals, response to superoxide, response to oxygen radical. The detailed data of these GO enriched terms are listed in Table 3, including enrichment FDR values and gene names of altered proteins associated with GO terms.   GO enrichment analysis of 20 significantly differentially expressed hippocampal proteins was carried out using the ShinyGO v0.74 tool (bioinformatics.sdstate.edu/go). The top thirty most significantly enriched GO terms for biological processes were summarized in hierarchical clustering tree, see Figure 3. The most enriched GO terms were related to metabolic changes: phosphorus metabolic process, phosphate-containing compound metabolic process, ATP metabolic process, methylglyoxal metabolic process, generation of precursor metabolites and energy, nucleoside phosphate metabolic process, nucleotide metabolic process; cytoskeleton organization: postsynaptic cytoskeleton organization, establishment of localization in cell, postsynaptic actin cytoskeleton organization, and oxidative stress: removal of superoxide radicals, response to superoxide, response to oxygen radical. The detailed data of these GO enriched terms are listed in Table 3, including enrichment FDR values and gene names of altered proteins associated with GO terms.       The 83 phosphorylated protein spots were detected in the rat brain cortex, only three were significantly downregulated: tubulin beta-2A chain ↓2.4-fold, spot 1; tubulin alpha-1A chain ↓3.7-fold, spot 2 and actin, cytoplasmic 1 ↓2.7-fold, spot 3; Figure 4a, Table 1c. 3.2.1. Pro-Q ® Diamond Staining and Colloidal Coomassie Blue Staining of 2D Gels The 83 phosphorylated protein spots were detected in the rat brain cortex, only three were significantly downregulated: tubulin beta-2A chain ↓2.4-fold, spot 1; tubulin alpha-1A chain ↓3.7-fold, spot 2 and actin, cytoplasmic 1 ↓2.7-fold, spot 3; Figure 4a, Table 1c. CBB staining revealed 87 protein spots in cortical 2D gels, only three of these were significantly altered. nLC-MS/MS analysis identified two proteins with decreased level (calmodulin 1 ↓2.0-fold, spot 1; alpha-synuclein ↓1.4-fold, spot 2) and one upregulated protein: mitochondrial isocitrate dehydrogenase [NAD] subunit alpha ↑2.4-fold, spot 3; Figure 4b, Table 1d.

GO Enrichment Analysis of Altered Cortical Proteins
GO enrichment analysis of six significantly differentially expressed cortical proteins (Table 1c,d) was carried out using the ShinyGO v0.74 tool (bioinformatics.sdstate.edu/go). The top ten most significantly enriched GO terms for biological processes were summarized in hierarchical clustering tree ( Figure 5). The most enriched GO terms were related CBB staining revealed 87 protein spots in cortical 2D gels, only three of these were significantly altered. nLC-MS/MS analysis identified two proteins with decreased level (calmodulin 1 ↓2.0-fold, spot 1; alpha-synuclein ↓1.4-fold, spot 2) and one upregulated protein: mitochondrial isocitrate dehydrogenase [NAD] subunit alpha ↑2.4-fold, spot 3; Figure 4b, Table 1d.

GO Enrichment Analysis of Altered Cortical Proteins
GO enrichment analysis of six significantly differentially expressed cortical proteins (Table 1c,d) was carried out using the ShinyGO v0.74 tool (bioinformatics.sdstate.edu/go). The top ten most significantly enriched GO terms for biological processes were summarized in hierarchical clustering tree ( Figure 5). The most enriched GO terms were related to vesicle endocytosis: presynaptic endocytosis, synaptic vesicle endocytosis, synaptic vesicle recycling, synaptic vesicle cycle, vesicle-mediated transport in synapse, and regulation of catecholamine uptake. The detailed data of these GO enriched terms are listed in Table 4, including enrichment FDR values and gene names of altered proteins associated with GO terms. to vesicle endocytosis: presynaptic endocytosis, synaptic vesicle endocytosis, synaptic vesicle recycling, synaptic vesicle cycle, vesicle-mediated transport in synapse, and regulation of catecholamine uptake. The detailed data of these GO enriched terms are listed in Table 4, including enrichment FDR values and gene names of altered proteins associated with GO terms.
According to the current annotations (https://www.uniprot.org) in the UniProt database, the identified cerebellar proteins with changed expression level were involved in metabolism (4), signal transduction (3), brain development (3), cytoskeleton organization (2), aging (2), protein transport (1), protein folding (1), RNA processing (1), and apoptosis (1) (Figure 7d; Table 6b). Table 5. nLC-MS/MS analysis of significantly altered protein spots identified in the rat striatum (a) and cerebellum (b) isolated from animals after 6 months of morphine withdrawal.     Figure 6. Representative 2D gel maps of total protein profiles in the rat striatum (a, upper panels) and cerebellum (b, lower panels) isolated from animals after 6 months of morphine withdrawal. Red arrows and numbers show the significantly altered protein spots.

Discussion
During the last years, we applied 2D electrophoresis and label-free quantification to find the significant alterations in protein expression in the rat brain cortex and hippocampus after chronic morphine treatment (10-50 mg/kg, 10 days) followed by different withdrawal periods (3 weeks, 3 months, 6 months) [15][16][17]28]. The aim of this work was to analyze the spectrum of altered proteins in selected rat brain regions after 6-month morphine withdrawal. Our study could have two main limitations. First, some animals could be lost during morphine administration and withdrawal. For that reason, we had three extra animals in each testing group. Second, proteomic analyses require a large amount of tissue. For this purpose, we selected the stated brain regions (cortex, hippocampus, striatum, and cerebellum). This selection provided us with a relatively large amount of biological material.
2D-DIGE analysis of the rat hippocampus showed that 10-day morphine administration results in a significant change of six proteins functionally related to metabolism, cytoskeleton organization, neuronal plasticity, apoptosis, and oxidative stress. Interestingly, the number of differentially regulated proteins was increased to 13 after 3 weeks of drug abstinence. Moreover, the level of α-synuclein (Snca), β-synuclein (Sncb), α-enolase (Eno1), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) persisted altered for 3 weeks since the withdrawal of morphine [16]. Ten proteins were identified in hippocampal 2D CBB-stained gels 3 months after cessation of 10-day morphine treatment; 14 proteins were significantly hypophosphorylated [17]. Among these, several glycolytic enzymes, such as GAPDH, Eno1, phosphoglycerate mutase 1 (Pgam1), triosephosphate isomerase (Tpi1) and fructose-bisphosphate aldolase A (Aldoa) were decreased. In this work, the impact of 6-month morphine withdrawal on the change of total protein composition was higher-14 proteins were identified (Figure 1b; Table 1b). On the other hand, the number of dysregulated phosphoproteins was reduced from 14 to 6 ( Figure 1a; Table 1a). The amount of glycolytic enzymes with significantly changed expression level was not so eminent. Pro-Q ® Diamond staining revealed downregulation of Aldoa and an increased level of Tpi1 was detected by CBB protein staining.
Previous studies of other authors have shown that enzymes involved in cellular metabolisms, such as glycolysis and Krebs cycle were altered in opioid-abusing patients and animal models [33,34]. Among the glycolytic enzymes, GAPDH is of particular interest. Its post-translational modifications may contribute to numerous cellular functions, including intracellular transport, cytoskeleton plasticity, heme chaperoning, transcription, and apoptosis [18,[35][36][37][38][39][40]. However, its role in apoptosis is not clear. Some studies describe its proapoptotic function, others a protective role [41]. One of the typical features of GAPDH is its use as a loading marker in hundreds of studies. Notably, it was shown that the quantity of GAPDH can vary under stressful conditions [42].
Decreased level of superoxide dismutase [Cu-Zn] (Sod1) was detected in our hippocampal samples after 3 weeks of drug abstinence [16]. In this work, 6-month morphine withdrawal revealed downregulation of peroxiredoxin-2 (Prdx2) and upregulation of Parkinson disease protein 7 homolog (Park7), see Table 1b. Park 7 is involved in the protection against oxidative stress [43]. Due to its protective role, Park 7 represents an ideal possible therapeutic target for Parkinson's disease (PD) and neurodegeneration [44]. We may hypothesize that stress-related pathways become activated during opioid withdrawal and can persist for several months after cessation of morphine administration.
Morphine could participate in the development of oxidative stress by promoting the formation of free radicals or reducing the activity of the antioxidant defense system which maintain redox homeostasis. Both these ways of action can be possibly combined [45]. Among the most important molecules playing a crucial role in cell protection against oxidative damage belong enzymes such as superoxide dismutase, glutathione peroxidase, catalase, and tripeptide glutathione [46]. The activity of antioxidant enzymes is closely associated with the production of reactive oxygen species (ROS) and reactive nitrogen species (RNS) that may lead to oxidative damage of DNA, lipids and proteins [47,48]. Results obtained by Motaghinejad et al. [49] showed that subcutaneous injection of morphine to rats significantly increased lipid peroxidation and decreased the activities of superoxide dismutase and glutathione peroxidase. Abdel-Zaher et al. [50] reported that glutamate levels and lipid peroxide malondialdehyde levels were significantly increased in the brain of morphine-treated mice. The impact of morphine on cellular redox balance may depend on multiple factors, such as species, age and sex of an organism, type of tissue, dosage, and length of usage [51].
The number of altered phosphorylated cytoskeletal proteins in the rat hippocampus was decreased from four (hypophosphorylation of F-actin-capping subunit beta (Capzb); actin, cytoplasmic 2 (Actg1); glial fibrillary acidic protein (GFAP) and tubulin alpha-1A chain (Tuba1a)) to one hyperphosphorylated tubulin alpha-1B chain (Tuba1b) when compared the period of abstinence 3 and 6 months. The change in protein expression was almost similar after 3 or 6 months of drug withdrawal and resulted in a decreased level of tubulin beta-4B chain (Tubb4b), tubulin polymerization promoting protein (Tppp), actin, cytoplasmic 1 (Actb), tubulin beta chain (Tubb4b), and alpha-internexin (Ina). These findings suggest that protracted morphine abstinence may cause long-term homeostatic changes in hippocampal plasticity [52]. However, according to our unpublished behavioral studies, we did not find significant differences between morphine-withdrawn and control animals. We may speculate that proteomic changes in the rat hippocampus after 6 months of morphine withdrawal do not require alterations in a certain behavior or functional state.
Twenty-eight significantly altered proteins were detected in the rat brain cortex after treatment with morphine for 10 days, this amount was reduced to 14 proteins after 3 weeks of abstinence [15], to 10 proteins after 3-month morphine withdrawal [17] and to only three proteins after 6-month drug abstinence. Chronic morphine treatment resulted in the decreased level of several glycolytic enzymes, such as Tpi1, Pgam1, GAPDH, Aldoa, pyruvate kinase PKM (Pkm), and phosphoglycerate kinase 1 (Pgk1). The expression level of Pkm, Pgk1 and GAPDH persisted decreased for 3 months of drug withdrawal and was not altered after 6 months of abstinence. It would be useful to confirm whether proteomicsindicated alterations in enzyme levels reflect changes in their activity. The change in protein expression does not necessarily mean the change in functional activity, as described by Bodzon-Kulakowska et al. [53] and Antolak et al. [26]. We assume that simultaneous alterations in both features may represent new insight into brain energy homeostasis.
Increasing evidence suggests that the striatum and cerebellum participate in drug addiction [54][55][56][57][58]. As a complement to proteomic analyses of rat hippocampus and cortex, we also performed the screening of protein alterations in the rat striatum and cerebellum after 6-month morphine withdrawal. Interestingly, the number of altered proteins was increased in both the striatum and cerebellum after 6-month drug withdrawal in comparison with the effect of 3-month morphine abstinence. In the striatum, the number of differentially expressed proteins was increased from 7 to 10 while in the cerebellum from 4 to 11. The majority of changes were related to metabolic alterations (L-lactate dehydrogenase B chain (Ldhb), malate dehydrogenase (Ldh1), creatine kinase B-type (Ckb), fructose-bisphosphate aldolase C (Aldoc), and Tpi1) ( Figure 6; Table 6a,b). Taken together, our results suggest that protracted morphine withdrawal causes significant proteomic changes in the energy metabolism of different rat brain parts. We assume that deeper metabolic investigation into the brain structures may reveal numerous differences in glucose metabolism, the tricarboxylic acid cycle (TCA) and fatty acid metabolism. In addition, we may expect changes in metabolites related to antioxidant and nucleotide pathways. However, detailed studies are missing.

Conclusions
Our data show that the rat hippocampus is more affected than the forebrain cortex in both protein phosphorylation and protein expression by 6-month morphine withdrawal. Gene Ontology (GO) enrichment analysis for 20 up-and downregulated proteins in the hippocampus revealed that the most enriched GO terms were associated with alterations in energy metabolism, cytoskeleton organization, and oxidative stress response. Our previous proteomic studies indicated that 10-day morphine administration results in significant alterations related to energy metabolism. Moreover, these changes persisted several weeks/months after the cessation of 10-day morphine treatment. We hypothesize that alterations in energy metabolism may be one of the functional consequences of the impaired antioxidant defense system. However, these questions need further investigation.