Proteomic analysis of protein composition of rat hippocampus exposed to morphine for 10 days; comparison with animals after 20 days of morphine withdrawal

Opioid addiction is recognized as a chronic relapsing brain disease resulting from repeated exposure to opioid drugs. Cellular and molecular mechanisms underlying the ability of organism to return back to the physiological norm after cessation of drug supply are not fully understood. The aim of this work was to extend our previous studies of morphine-induced alteration of rat forebrain cortex protein composition to the hippocampus. Rats were exposed to morphine for 10 days and sacrificed 24 h (groups +M10 and −M10) or 20 days after the last dose of morphine (groups +M10/−M20 and −M10/−M20). The six altered proteins (≥2-fold) were identified in group (+M10) when compared with group (−M10) by two-dimensional fluorescence difference gel electrophoresis (2D-DIGE). The number of differentially expressed proteins was increased to thirteen after 20 days of the drug withdrawal. Noticeably, the altered level of α-synuclein, β-synuclein, α-enolase and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was also determined in both (±M10) and (±M10/−M20) samples of hippocampus. Immunoblot analysis of 2D gels by specific antibodies oriented against α/β-synucleins and GAPDH confirmed the data obtained by 2D-DIGE analysis. Label-free quantification identified nineteen differentially expressed proteins in group (+M10) when compared with group (−M10). After 20 days of morphine withdrawal (±M10/−M20), the number of altered proteins was increased to twenty. We conclude that the morphine-induced alteration of protein composition in rat hippocampus after cessation of drug supply proceeds in a different manner when compared with the forebrain cortex. In forebrain cortex, the total number of altered proteins was decreased after 20 days without morphine, whilst in hippocampus, it was increased.


Introduction
Opioid dependence and withdrawal syndrome are the leading problems associated with licit and illicit opioid use. In spite of the large literature on opioid addiction, there is only little available information about the durability and reversibility of potentially adverse effects of opioid treatment and withdrawal. From a clinical point of view, drug withdrawal is the leading pathophysiological state driving opioid dependence and addictive behaviors [1,2]. Here, we focused our attention on morphine as the prototypical opioid agonist with which all others are compared.
Our previous results, in accordance with the data of Sim et al. [3], Sim-Selley et al. [4] and Maher et al. [5], indicated a desensitization of (2-D-alanine2-4-methylphenylalanine-5-glycineol)-enkephalin (DAMGO)-and (2-D-alanine-5-D-leucine)-enkephalin (DADLE)-stimulated G-protein responses in plasma membranes (PM) isolated from forebrain cortex (FBC) of rats exposed to morphine for 10 days [6]. Behavioral tests proved that these rats developed a tolerance to this drug as there was no significant difference between control (−M10) and morphine-treated (+M10) rats in the sensitivity to heat stimulation (hot-plate test), which was determined as a delay in hind paw withdrawal. Tolerance to morphine in (+M10) rats was also evidenced by hind paw withdrawal test. Precipitation of morphine withdrawal state by naloxone resulted in a rapid and dramatic opiate abstinence syndrome. There were no detectable signs of abstinence syndrome (such as body shakes, teeth clatter, vacuous chewing, ptosis, irritability to touch, diarrhea), in the (−M10) animals.
In opioid-dependent subjects, the body has adapted to perpetually high opioid tone by making homeostatic adjustments in anti-opioid systems [7]. In an effort to clarify biochemical mechanisms of development of opiate tolerance and dependence, our previous results showed that in rat FBC, such homeostatic adjustments were accompanied by a reversible and specific up-regulation of adenylyl cyclases I and II (ACI and ACII). Importantly, the up-regulation of ACI and ACII disappeared after 20 days of morphine withdrawal (±M10/−M20) [8,9].
Proteomic analyses of the consequences of a 10-day morphine treatment and subsequent 20-day drug withdrawal on FBC, which were based on CBB-staining of 2D gels and matrixassisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS/MS), showed that the number of altered proteins was decreased from 28 (determined in ±M10 rats) to 14 (determined in ±M10/−M20 rats). When using label-free quantification (LFQ), the number of altered proteins was decreased from 113 to 19 [10]. Thus, we brought a straightforward evidence for the ability of the organism to oppose the drastic, morphine-induced change of the target tissue protein composition with the aim to return to the physiological norm after a complete removal of the drug.
The chronic administration of opioid drugs was reported to modulate synaptic transmission and plasticity of hippocampus and inhibit adult neurogenesis [11]. Chronic opioid administration also resulted in obtuseness of spatial memory and increase in the expression of proteins functionally associated with apoptosis and neurotoxicity [12]. In contrast to chronic opioid treatment, opioid withdrawal was associated with enhanced hippocampal plasticity [13]. Interestingly in the context with studies of prolonged morphine effect on the brain, the chronic antidepressant treatment was found to increase neurogenesis in adult rat hippocampus [14].
As neurodegeneration was frequently described as a pathological state accompanying addiction to opioid drugs and hippocampus is regarded as one of the key brain areas [15,16], together with striatum [17] and cerebellum [18], which exhibit neurogenesis in adult brain, the aim of this work was to extend our previous proteomic studies of morphine-induced alteration of rat forebrain cortex protein composition to the hippocampus.
In the first set of experiments, samples from (+M10) group were labeled with Cy3 and samples from control group (−M10) were labeled with Cy5. In the second set of experiments, samples prepared from rats sacrificed 20 days after morphine withdrawal (+M10/−M20) were labeled with Cy3; samples from the control group (−M10/−M20) were labeled with Cy5. Cy2 was always used to label a mixture of equal amounts of protein taken from all samples of PNS. Cy2-labeled protein mix thus represented an internal standard.
Samples were briefly vortexed and incubated on ice for 30 min in the dark. Staining was terminated by the addition of 10 mM lysine for 10 min on ice in the dark. Dye-labeled PNS samples were combined in such a way that each mixture was comprised of protein samples from (+M10, −M10) or (+M10/−M20, −M10/−M20) groups plus the aliquot of the internal standard (1:1:1, v/v/v). Finally, the mixture of labeled samples was mixed with an equal amount of sample buffer (8 M urea, 130 mM dithiothreitol (DTT), 4% (w/v) CHAPS, 2% (v/v) BioLyte 3-10 buffer (Bio-Rad).
All the samples (250 μl) were then transferred into a groove of the Immobiline DryStrip Reswelling Tray (GE Healthcare). Immobiline DryStrips (linear pH gradient 3-11 NL, 13 cm) were placed into the protein samples and rehydrated overnight. The exposure of protein labeled with CyDyes to all light sources was kept to a minimum. Isoelectric focusing was performed using the Multiphor II system (GE Healthcare) at 14˚C in the following manner: 150 V for 5 h, 500 V for 1 h, 3500 V for 12 h and 500 V for 3 h. The focused strips were stored at-20˚C or immediately used.
The strips were rinsed thoroughly with ultrapure water, dried quickly on filter paper and equilibrated in 5 ml of equilibration buffer (50 mM Tris-HCl pH 6.8, 6 M urea, 0.1 mM EDTA, 2% SDS, 30% glycerol and 0.01% bromophenol blue) containing 1% DTT for 10 min in order to reduce disulfide bridges and other oxidized groups. Subsequently, the strips were alkylated in equilibration buffer containing 2.5% iodoacetamide for 10 min. Molecular weight markers were loaded onto a piece of filter paper and placed close to the alkaline side of the strip. The strip and molecular marker were covered with 0.5% agarose. Gels were run vertically at a constant current of 20 mA for 20 min and then at 90 mA for 4 h till the bromophenol blue dye reached the end of the gel. The apparatus was cooled to 15˚C using the Hoefer SE 600 unit (GE Healthcare). All the 2D-DIGE analyses were performed three times. After electrophoresis, the gels were washed with ultrapure water for 2×15 min before scanning.

2D image analysis
Three different gel images were obtained from one gel at the appropriate wavelength. They are Cy2 (blue 488 nm laser and 520 nm band pass emission filter), Cy3 (green 532 nm laser and 580 nm band pass emission filter) and Cy5 (red 633 nm laser and 670 nm band pass emission filter) by using a Pharos FX™ scanner (Bio-Rad) at a resolution of 50 μm.

Colloidal Coomassie staining
For MS analysis, the gels were stained by colloidal Coomassie brilliant blue G-250 (CBB) to enable the visual detection according to Fountoulakis et al. [23]. The fresh 2D gels were immediately fixed in 50% methanol/7% acetic acid for 1 h and incubated with CBB (17% ammonium sulfate, 34% methanol, 3% orthophosphoric acid and 0.1% CBB) overnight with gentle agitation. After staining, the gels were washed several times in ultrapure sterile water and stored in 1% acetic acid at 4˚C.

MALDI-TOF MS/MS
Selected spots with significantly changed expression (�2-fold) were cut out from 2D gels and processed as described previously [10,22]. Briefly, chopped 1x1x1 mm pieces were covered with 100 μl of 50 mM ammonium bicarbonate (ABC) buffer in 50% acetonitrile (ACN) (buffer A) with 50 mM DTT. After sonication, the supernatant was removed and each gel spot was mixed with 100 μl of buffer A with 50 mM iodoacetamide (IAA). After sonication, the supernatant was discarded and replaced with 100 μl of buffer A with 50 mM DTT. After sonication, the supernatant was discarded, and the samples were sonicated for 5 min in 100 μl of HPLC/MS-grade water. The water was then discarded, and the samples were again sonicated for 5 min in 100 μl of ACN. The ACN was removed, 5 ng of trypsin in 10 μl of 50 mM ABC was added to each sample, and the samples were incubated at 37˚C overnight. Trifluoroacetic acid (TFA) and ACN were added to reach final concentration of 1% TFA and 30% ACN. After sonication, 0.5 μl aliquot of trypsin digest was transferred onto MALDI target and let to dry. Subsequently, 0.5 μl drop of alpha-cyano-4-hydroxycinnamic acid solution (10 mg/ml in 50% ACN) was transferred onto MALDI target and let to dry again. Samples were measured using a 4800 Plus MALDI TOF/TOF analyzer (Applied Biosystems/MDS Sciex) equipped with a Nd:YAG laser (355 nm, firing rate 200 Hz).
The data were analyzed using in house running Mascot server 2.2.07 and matched against comprehensive UniProt database of protein sequences (27929 sequences; 14725510 residues). Database search criteria were as follows: enzyme = trypsin; taxonomy = Rattus norvegicus. Cystein carbamidomethylation was set as fixed modification, methionine oxidation and deamidation as variable modifications, respectively. Peptide mass tolerance was set to ±100 ppm, and fragment mass tolerance to ±0.4 Da with a maximum of two missed cleavages. Only hits that were scored as significant (MASCOT score �57, p<0.05) were accepted. All MALDI-TOF MS/MS analyses were performed in duplicates.
All data were analyzed and quantified with MaxQuant software. The false discovery rate (FDR) was set to 1% for both proteins and peptides and a minimum length of seven amino acids was specified. The Andromeda search engine was used for the MS/MS spectra search against the UniProt Rattus norvegicus database of protein sequences. Enzyme specificity was set as C-terminal to Arg and Lys, also allowing cleavage at proline bonds and a maximum of two missed cleavages. Dithiomethylation of cysteine was selected as fixed modification and Nterminal protein acetylation and methionine oxidation as variable modifications. The "match between runs" feature of MaxQuant was used to transfer identifications to other LC-MS/MS runs based on their masses and retention time (maximum deviation 0.7 min) and this was also used in quantification experiments. Quantifications were performed with the label-free algorithms according to Cox et al. [21]. Binary logarithms of intensity ratios were then median calculated for each group and the difference between control and sample was determined. Only at least 1.7-fold significant differences calculated for at least 2 measured values from 3 replicates were taken into consideration.

Protein determination
Lowry method was used for determination of protein concentration in all hippocampal samples of PNS (+M10, −M10, +M10/−M20 and −M10/−M20) using bovine serum albumin (Sigma, Fraction V) as a standard. Data were calculated by fitting the calibration curve as a quadratic equation.

Results
Comparative proteomic analysis of rat hippocampus from animals exposed to morphine for 10 days (+M10) and from control animals (−M10) by
The presence of the same protein in more than just one spot, when resolved by 2D-ELFO, represents according to our previous experience [10,22], evidence for the existence of different subunits of the same protein. In general, this multiplicity occurs due to post-translational modifications of such proteins which result in the production of multiple spots when resolved in 2D gels [25].
Comparative proteomic analysis of rat hippocampus from animals exposed to morphine for 10 days and sacrificed 20 days since the last dose of morphine (+M10/−M20) and from control animals (−M10/−M20) by

2D-DIGE and MALDI-TOF MS/MS
Samples of PNS were labeled with Cy5 (−M10/−M20) or Cy3 (+M10/−M20) and resolved by 2D-ELFO as described in Methods together with internal standard labeled with Cy2 (Fig 4). and morphine-treated rats (+M10) exposed to increasing doses of morphine for 10 days. PNS fractions were prepared as described in Methods. Staining of PNS samples and resolution of Cy5 (−M10)-, Cy3 (+M10)-and Cy2 (mixed)-labeled proteins in 2D gels (first dimension on pH 3-11 IPG strips and then by SDS-PADE on 10% acrylamide gels) was performed as described in Methods. White arrows indicate the protein spots altered at least 2-fold by morphine (p�0.05). Comparison of DIGE gels was done by PDQuest™ (Bio-Rad). The significance of the difference between the three sets of (−M10) and (+M10) gels was determined by Student´s t-test. The positions of molecular weight markers are indicated on the right side and pI at the bottom of each gel.

Immunoblot analysis of αand β-synuclein, GAPDH and actin in rat hippocampus of experimental groups (±M10) and (±M10/−M20)
In order to verify the results obtained by comparative 2D-DIGE, PNS hippocampal fractions prepared from groups (−M10), (+M10), (+M10/−M20) and (−M10/−M20) were resolved by 2D-ELFO and the content of α-and β-synuclein, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and actin was determined by immunoblotting with specific antibodies. In (±M10) samples of PNS, position of immunoblot signals of α-and β-synuclein in 2D gels (Fig 6A) was the same as in 2D-DIGE gels [compare with Fig 1 and Table 1; spot 1 (βsynuclein, #2.0-fold) and spot 2 (α-synuclein, #2.5-fold)]. The M w of these two proteins was � 14 and 16 kDa, and their pI values were � 4.3 and � 4.7, respectively. Intensities of immunoblot signals were decreased to 41.3% (β-synuclein, p<0.01) and 46.6% (α-synuclein, p<0.01) when compared with corresponding controls. Thus, results obtained by 2D-DIGE indicating  Table 3. the decrease of these two proteins by morphine-treatment, were verified by immunoblot analysis of 2D gels. The difference in the magnitude of the decrease may be attributed to the major difference in methodology of detection combined with the necessity to apply largely different amounts of total protein per gel (75 μg versus 2000 μg) to achieve the proper labeling of the sample with Cy probes on the one hand and intensity of immunoblot signals on the other hand [26,27]. Accordingly, the densitometric scanning of immunoblot signals of α/β-synucleins in (±M10/−M20) samples of PNS indicated a significant down-regulation to 58% (β-synuclein, p<0.01) and 85% (α-synuclein, p>0.05), respectively (Fig 6B). When combined together, the results of both 2D-DIGE and immunoblot analyses indicated consistently the down-regulation of α-and β-synuclein in hippocampus of rats exposed to morphine for 10 days. Noticeably, this down-regulation persisted for 20 days of abstinence.
Similar results were obtained in (±M10/−M20) samples of PNS. GAPDH signals were distributed over a wide range of pI � 3.5-10, M w of all isoforms of this enzyme was � 36 kDa and morphine-treatment resulted in the transfer from alkaline to acid region of the gel. However, the relative magnitude of this transfer was less than in (±M10) samples of PNS (Fig 7B, lower  panels). Intensities of individual spots in alkaline region (blue arrows) were decreased to � 58% (spot 3, p<0.01), � 60% (spot 4, p<0.01), � 64% (spot 5, p<0.01) and � 36% (spot 6, p<0.05); the two minor spots (1 and 2) were increased to � 151% (p<0.05) and � 145% (spot 2, p<0.05), respectively. The major spots 3, 4 and 5 were those identified in DIGE gels as spots 11, 12 and 13 (red arrows; please compare with Figs 4 and 2B and Table 3). The total signal of GAPDH subunits in alkaline region (1+2+3+4+5+6) was decreased to 72% (p<0.05); the total signal of all spots was unchanged by morphine (NS, p>0.05). The total signal of actin was unchanged, Fig 8. In (±M10) samples of PNS (Fig 8A), the major signal of actin was decreased to 80%, but this decrease was not significant when averaged in three immunoblots (NS, p>0.05); the minor signal was increased to 152% ( � , p<0.05). The total signal of both spots was unchanged (NS, p>0.05). The same result was obtained in (±M10/ −M20) groups of rats (Fig 8B): the major signal of actin was unchanged when averaged in three blots (NS, p>0.05), the minor signal was increased to 173% ( � , p<0.05). The total signal of both spots was unchanged (NS, p>0.05). As actin is known to be modified by sugars, the increased  Table 1). In (B), the immunoblot signals 3, 4 and 5 were identical with spots 11, 12 and 13, which were recognized as significantly decreased (�2-fold) by DIGE analysis (compare with Figs 4 and 2B and Table 3). Data present bar graphs with the individual data points along with the bar with error. The significance of the difference between the (±M10) or (±M10/−M20) samples of PNS was analyzed by Student´s t-test using GraphPadPrism4; � , p<0.05; �� , p<0.01. https://doi.org/10.1371/journal.pone.0231721.g007

PLOS ONE
level of its minor fragment exhibiting the higher molecular weight may represent the glycosylated form detected in both (±M10) and (±M10/−M20) samples of hippocampus.
The level of beta-actin commonly used as a loading control was also detected and was unchanged (NS, p>0.05), S1 File. Moreover, 1D immunoblot resolution of GAPDH was performed (S2 File). The difference in the expression level between (±M10) and (±M10/−M20) samples was not significant.

Discussion
2D-DIGE and 2D immunoblot analysis of rat hippocampus from animals exposed to morphine for 10 days (±M10); comparison with animals exposed to morphine and subsequently nurtured for 20 days in the absence of this drug (±M10/−M20) The number of altered proteins in hippocampus was increased from 6 to 13 after 20 days of abstinence. This finding is just the opposite when compared with that observed in forebrain cortex, where the number of differentially expressed proteins was decreased from 28 (±M10) to 14 (±M10/−M20) when determined in CBB-stained 2D gels or from 113 to 19 when determined by LFQ [10]. Noticeably, the altered level of four proteins identified in (±M10) samples of PNS (β-synuclein, α-synuclein, alpha-enolase and GAPDH) persisted for 20 days since the withdrawal of morphine (Figs 1 and 4, Tables 1 and 3).
The α-synuclein is a pathological protein functionally related to production of "Lewy bodies" in Parkinson´s disease and dementia [28][29][30][31]. While its aggregation represents the major risk factor for neurodegeneration, its function under physiological conditions remains poorly understood. It has been suggested that α-synuclein is involved in negative regulation of dopaminergic neurotransmission, synaptic vesicle cycling, synaptic plasticity and neuroprotection [32]. Interestingly, Ziolkowska et al. [33] observed the down-regulation of α-synuclein mRNA in the basolateral amygdala, dorsal striatum, nucleus accumbens and ventral tegmental area of mice withdrawn from chronic morphine treatment for 48 h; on the other hand, protein level of α-synuclein, determined in the same brain regions, was up-regulated. β-synuclein was described as the non-amyloidogenic homolog of α-synuclein with anti-apoptotic effect [34]. The evidence for down-regulation of α/β synucleins persisting for 20 days since morphine withdrawal was confirmed by 2D immunoblot analysis, Fig 6. In 2010, the group of Dr. Piotr Suder created the Morphinome Database (www.addictionproteomics.org) in order to facilitate the search for the proteins altered by morphine administration. This database is continuously updated. More than 20% of proteins distinguished in this database are functionally related to regulation of energy metabolism-glycolysis, gluconeogenesis, Krebs cycle and oxidative phosphorylation [35].
In our study, the down-regulation of glycolytic enzyme GAPDH was observed after 10 days of morphine treatment and this decrease was also noticed in hippocampus of rats sacrificed after 20 days of drug withdrawal. Noticeably, the 2D immunoblot analysis recognized the subset of six GAPDH spots (in both ±M10 and ±M10/−M20 samples of PNS) which were present in alkaline region of 2D gels and selectively decreased by morphine (Fig 7). As GAPDH was found to be extensively modified by post-translational modifications [36][37][38][39][40], and post-translational modifications of this enzyme were suggested to play a role in oxidative stress and apoptosis [41,42], it is reasonable to assume that morphine-induced change of six GAPDH subunits reveals different functions fulfilled by individual variants of this enzyme. One of the functional consequences of manifestation of oxidative stress is the decrease of cellular ATP level and blockade of glycolysis [43]. Accordingly, proteomic analysis of hippocampus from mice exposed to morphine for 10 days indicated a decreased level of three enzymes related to regulation of glycolysis and mitochondria: E2 component of the pyruvate dehydrogenase complex, lactate dehydrogenase 2, and Fe-S protein 1 of NADH dehydrogenase [44]. Further view on literature data dealing with morphine-induced alteration of CNS energy metabolism suggests that chronic morphine treatment impairs the glucose metabolism and ATP production. This effect is associated with morphine withdrawal symptoms and impairment in memory [45].
Our data also indicated an increased level of α-enolase, another glycolytic enzyme participating in a multitude of pathological processes. Increased level of α-enolase was again noticed in (±M10) as well as (±M10/−M20) samples of hippocampus (Tables 1 and 3). Based on comparative proteomic analysis of human, mouse and rat tissues, Díaz-Ramos et al. [46] reported that α-enolase could be regarded as a marker of "pathological stress" proceeding in a range of diseases-cancer, skeletal myogenesis, Alzheimer´s disease, rheumatoid arthritis, inflammatory bowel disease, autoimmune hepatitis and membranous glomerulonephritis. Morphine administration resulted in up-regulation of α-enolase in striatal neuronal cell cultures [47], but the change of this protein was not detected in hippocampus of morphine-treated rats by proteomic analysis of Bierczynska-Krzysik et al. [48].
Apart from protein changes persisting for 20 days after morphine withdrawal, decreased levels of tubulin α-1A chain, cofilin-1 and superoxide dismutase were found in PNS prepared from (±M10/−M20) rats ( Table 3). Data of Marie-Claire et al. [49] indicated decreased mRNA and protein level of α-tubulin in rat striatum after chronic morphine treatment, whilst proteomic analysis of Bodzon-Kulakowska et al. [47] indicated the increase of α-and β-chains of tubulin in striatal neuronal cell cultures after morphine exposure. The full understanding of the functional meaning of down-or up-regulation of cytoskeleton protein tubulin α-1A will therefore require further attention.
The large decrease of superoxide dismutase (#3.1-fold) in hippocampal samples (±M10/ −M20) ( Table 3) may be interpreted as the evidence of attenuated efficiency of hippocampus recovering from prolonged morphine effect in protection against oxidative stress. Superoxide dismutase, the enzyme catalyzing the conversion of superoxide radical O 2 ─ to O 2 and H 2 O 2 , represents an important antioxidant system protecting mammalian cells against oxidative damage [50,51]. This protection is necessary as morphine treatment was reported to modulate both enzymatic and non-enzymatic antioxidant defense systems and induce oxidative stress and apoptosis under in vivo conditions [52,53].
Label-free quantification of rat hippocampus from animals exposed to morphine for 10 days (±M10); comparison with animals exposed to morphine and subsequently nurtured for 20 days in the absence of this drug (±M10/−M20) With the aim to obtain a larger coverage of proteome changes elicited by morphine, we have also used the LFQ technique. The results supported those obtained by gel-based proteomic analysis. The LFQ identified 19 proteins with significantly changed expression level (�1.7-fold) in (±M10) samples and this number was increased to 20 in (±M10/−M20) samples of hippocampus, Tables 5 and 6, Fig 9. Thus, the lack of reversibility of chronic morphine effect after 20 days of drug withdrawal was confirmed. What type of underlying mechanism could cause this lack of reversibility? Opioids were reported to cause significant changes of glutamatergic transmission [54], neurogenesis [55], dendritic stability [56] and long-term potentiation [57][58][59]. Data of Cai et al. [60] indicated that morphine-treatment of mice for 6 days resulted in a decrease of excitatory synapse densities in parallel with an enhancement of densities of inhibitory synapses in hippocampus. This effect was mediated by μ-opioid receptors. In primary hippocampal neurons, the short-term exposure to morphine (0.5-24 h) induced an increase of intracellular levels of reactive oxygen species (ROS) generated by NADPH-oxidase. The increase of ROS was followed by an increase of endoplasmic reticulum stress markers and autophagy. Mattei et al. [12] suggested that i) hippocampus is a vulnerable part of brain susceptible to oxidative damage during aging and chronic stress, ii) prolonged withdrawal may lead to homeostatic changes heading for the restoration of the physiological norm and iii) abuse of morphine may lead to ROS-induced neurodegeneration and apoptosis.
The LFQ analysis of our withdrawal samples (Table 6) also confirmed the results of 2D-DIGE and 2D immunoblot analyses indicating the decrease of α-synuclein level (#1.7-fold). The α-synuclein was described to participate in negative regulation of apoptosis, aging and oxidative stress [61].
In humans, heroin abuse was shown to increase oxidative stress [62]-plasma concentrations of lipid peroxides (LPO) and nitric oxide were increased in parallel with a decrease of antioxidants, vitamin C, vitamin E and beta-carotene. In erythrocytes, LPO were increased but the activity of antioxidant enzymes superoxide dismutase, catalase and glutathione peroxidase was decreased; in mice [63], heroin treatment resulted in a decrease of total antioxidant capacity in serum and activity of antioxidant enzymes superoxide dismutase, catalase and glutathione peroxidase in brain. Disturbance of pro-antioxidant balance was associated with oxidative damage of brain DNA, proteins and lipids. Addition of exogenous antioxidants attenuated the oxidative stress.
Results presented in our work extend these data as the down-regulation of superoxide dismutase protein level (#3.1-fold) was detected in hippocampal samples collected from (±M10/ −M20) rats (Table 3). As discussed above, the large decrease in the expression level of this antioxidant enzyme may be interpreted as a piece of supportive evidence for attenuated efficiency of rat hippocampus recovering from addiction to morphine. Protection against oxidative damage is decreased and return to physiological norm obscured.

Conclusions
1. The 2D-DIGE analysis identified in the hippocampus of rats exposed to morphine for 10 days (+M10) six altered proteins when compared with samples prepared from control animals (−M10). In rats treated with morphine for 10 days and subsequently nurtured for 20 days in the absence of the drug (+M10/−M20), thirteen altered proteins were detected when compared with control animals (−M10/−M20). Thus, the number of protein spots with changed expression level (�2-fold) was increased 2-fold after 20 days of abstinence.
2. This result is just the opposite when compared with that detected in the forebrain cortex, where the number of altered proteins was decreased from 28 (±M10) to 14 (±M10/−M20) when determined in CBB-stained 2D gels or from 113 to 19 when determined by LFQ [10].

5.
When using LFQ analysis, nineteen proteins with significantly changed expression level (�1.7-fold) were identified in group (+M10) when compared with group (−M10) and the number of altered proteins was increased to twenty in hippocampal samples collected from rats after 20 days of morphine withdrawal (±M10/−M20). 6. We conclude that the morphine-induced alteration of protein composition in rat hippocampus after cessation of drug supply proceeds in a different manner when compared with the forebrain cortex. In the forebrain cortex, the total number of altered proteins was decreased after 20 days without morphine, whilst in the hippocampus, it was increased. Thus, the two functionally distinct parts of CNS respond to the disturbance of the homeostatic balance caused by drug addiction in a different manner with the aim to restore the physiological norm.
Supporting information S1 Table. MALDI-TOF MS/MS analysis of eight altered protein spots (with a complete list of peptides) in PNS prepared from hippocampus of rats exposed to morphine for 10 days