Male C57BL/6J mice at 12 weeks of age weighing 22–25 g were purchased from the Model Animal Research Center of Nanjing University and maintained on a 12-h light/dark cycle with controlled temperature (22–23°C) and humidity (45–55%). For all repertoires of experiments, C57BL/6J mice were randomly assigned to the Control, Chronic Alcohol Exposure (CAE), CAE + recombinant human erythropoietin (rhEPO) and CAE+rhEPO+all-trans-retinoic acid (ATRA) groups on average (8 mice/group). The CAE, CAE+rhEPO and CAE+rhEPO+ATRA groups were subjected to alcohol exposure for 4 weeks. Successively, the CAE+rhEPO and CAE+rhEPO+ATRA groups were treated with rhEPO in saline [100 IU/ml, 500 IU/kg body weight; Roche Diagnostics (Shanghai) Ltd., China] intranasally every 12 h for 4 weeks, while the CAE group was treated only with an equal amount of vehicle. Simultaneously, the CAE+rhEPO+ATRA group received intraperitoneal administration of ATRA in saline (2 mg/ml, 20 mg/kg; Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) 30 min before intranasal rhEPO treatment for 4 weeks, while other groups received merely an equal amount of vehicle. Animals in the Control group underwent all procedures except chronic alcohol exposure and medication. For neural function analyses, neurological deficits were assessed by motor cooperation, cognition and nerve conduction tests 12 h after rhEPO administration for the last time. For the following histopathological analyses, animals were decapitated and the brain specimens were dissected and subjected to immunostaining and electron microscopy. For the independent mechanism study, the homogenate of brain tissue of animals was subjected to western blotting analyses for neural damage repair-related key molecules 30 min immediately after the drug treatment for the last time. All experimental procedures on animals were approved by the Ethics Committee of Harbin Medical University and the Ethics Committee of Qiqihar Medical University, and followed the principles of medical ethics.

Chronic alcohol exposure was achieved via a chronic intermittent ethanol (CIE) vapor inhalation procedure previously described for C57BL/6J mice (24). Briefly, mice were placed in standard mouse cages in Plexiglas vapor chambers (60×36×60 cm, PlasLabs, up to six cages per chamber) and exposed to continuous fresh air-mixed vaporized which delivered 19–22 mg of alcohol per liter of air at a rate of ~10 l/min and produced blood alcohol levels of 175±25 mg/dl as designed. The blood alcohol level was verified weekly via blood samples taken from dedicated ‘sentinel’ mice exposed to alcohol simultaneously with the test mice. To initiate intoxication and stabilize blood alcohol levels, the alcohol groups received intraperitoneal injections of 1 mmol/kg body weight of the alcohol dehydrogenase inhibitor pyrazole (Sigma-Aldrich; Merck KGaA) combined with 1.5 g/kg body weight of 20% alcohol (v/v), in a volume of 10 ml/kg body weight, before placement in the chambers. Exposure lasted for 16 h each day (in at 17:00 h, 2 h before the start of the 12-h circadian dark phase, out at 09:00 h), followed by an 8-h withdrawal. There were 4 consecutive days of intermittent exposure (Monday through Thursday) and then an 80-h withdrawal (Friday through Monday). This complete weekly cycle was repeated four times. Controls received an injection of 1 mmol per kg of body weight pyrazole in saline to control for this treatment, before placement in dedicated air chambers (located adjacent to the alcohol chambers), which received vaporized air at the same exchange rate as the alcohol chambers.

The intranasal administration was performed as reported previously (25). In brief, recombinant human (rh) EPO [Roche Diagnostics (Shanghai) Ltd.] was prepared using sterile 0.45% normal saline solution at a concentration of 100 IU/ml and stored at −20°C. Each mouse was held ventral side up, a small-modified 27-French catheter was inserted into either nares, and saline or rhEPO was slowly administered intranasally at no more than 5 µl increments 5–10 min apart for a total of 500 IU/kg body weight. The mouse was held for 1–2 min to ensure absorption. Drug or saline was administered every 12 h beginning immediately after the 4-week alcohol exposure.

Motor cooperation of mice was assessed by rotarod and beam walking assays. For the motor tests, the training began 3 days before alcohol exposure and the tests were conducted 1 week after treatment. The rotarod assay was conducted using a ROTO-ROD series 8 (IITC Inc./Life Science, Inc., Woodland Hills, CA, USA) which could accelerate rotational speed from 4 to 40 rpm over 120 sec. The mean latency of mice to fall off the rotarod in the three trials was recorded as measurement. For the beam walking assay, the time mice spent in crossing a 1-m long and 3-cm thick beam was calculated for analysis.

Novel location recognition and open-field tests were used to evaluate the spatial hippocampal-dependent memory, activity and anxiety. For the novel location recognition test, each mouse was allowed to explore the empty arena (50-cm diameter) for 10 min on the 1st day for familiarization. In this arena, two identical objects were placed in the middle on the 2nd day for acquisition, and in the same arena, one of the two objects was moved to a novel location on the 3rd day for the test. The exploration time for the two objects in the acquisition and test phase was recorded. The percentage of the time mice spent exploring the displaced object in the total exploration time for both objects was defined as the preference index. For the open-field test, mice were placed into a chamber measuring 50×50×38 cm for 60 min. The total number of forelimb rears, and the ratio of time spent in the center to the time spent in the corners were recorded as measurements.

Somatosensory-evoked potential (SSEP) was monitored to assess cortical synaptic transmission and nerve conduction. To generate forepaw stimulation, 2 metal pins were inserted into the left forepaw and a 1–5 mA 1 or 2 msec electrical pulse was delivered by constant current stimulus output using the RM6240C Multi-channel Bioelectric Signal Acquisition System (Chengdu Instrument Factory, Sichuan, China). For electroencephalography (EEG) recording, a Teflon-coated silver chloride wire (125 µm) with an exposed tip (1 mm) as the EEG electrode was placed on the surface of the primary somatosensory cortex corresponding to the left forelimb within the agarose, and the reference electrode was placed on the nasal bone under the skin. The cortical signal was amplified, filtered and recorded using the RM6240C Multi-channel Bioelectric Signal Acquisition System. The onset latency and the amplitude of the large primary cortical wave were determined.

Mice were anesthetized with isoflurane and carbon dioxide as dictated by ethical guidelines and decapitated. The brains were dissected rapidly, fixed overnight in 4% paraformaldehyde and processed for paraffin embedding. For immunofluorescence staining, 4-µm paraffin sections were prepared. After deparaffinization and rehydration according to standard procedures, sections underwent antigen retrieval in citrate buffer (pH 6.0) using a microwave oven (750 W). Non-specific antibody binding was blocked with 5% bovine serum albumin (BSA), 0.2% Triton X-100 in PBS for 1 h at room temperature. Subsequently, sections underwent incubation with primary antibodies [rabbit polyclonal anti-human/mouse/rat synaptophysin (IgG; 1:100; catalog no. SAB4502906; Sigma-Aldrich; Merck KGaA) and rabbit polyclonal anti-human/mouse/rat myelin basic protein (MBP) (IgG; 1:100; catalog no. 78896; Cell Signaling Technology, Inc., Danvers, MA, USA)] diluted in 2% BSA, 0.1% Triton in PBS overnight at 4°C. Following the washing out of the unbound antibodies by PBS, the sections underwent incubation with the secondary antibody [goat polyclonal anti-rabbit IgGs conjugated to Alexa Fluor^® 488 (1:1,000; catalog no. A32723; Invitrogen, Thermo Fisher Scientific, Inc., Waltham, MA, USA)] diluted in 2% BSA, 0.1% Triton in PBS in dark for 1 h at room temperature, and counterstained with DAPI (1 µg/ml; Sigma-Aldrich; Merck KGaA) in methanol for nuclear staining. The stained sections were mounted with anti-fade reagent (Invitrogen, Thermo Fisher Scientific) and photographed under a BX51 fluorescence microscope (Olympus, Tokyo, Japan). The fluorescence intensity was measured using Image-Pro Plus 6.0 (Media Cybernetics, Inc., Rockville, MD, USA).

Samples of dissected brain were fixed in 0.1 M sodium-cacodylate-buffered mixture (5% glutaraldehyde and 4% paraformaldehyde), post-fixed in 1% osmium tetroxide and 1.5% potassium ferrocyanide, dehydrated and embedded in epoxy resin. Ultrathin sections (80 nm) were then cut, collected on copper grids, stained with lead citrate and uranyl acetate for contrast and examined under H-7650 electron microscope (Hitachi, Osaka, Japan) equipped with a MegaView III CCD camera (Soft Imaging System, Lakewood, CO, USA). The synapse densities in the captured images were quantified using Adobe Photoshop Creative Cloud.

The dissected brain tissue was homogenized and lysed using RIPA lysis buffer (Beyotime, Zhejiang, China) supplemented with protease and phosphatase inhibitors (Complete and Phosphostop, Roche, Upper Bavaria, Germany) for 30 min on mice, and centrifugation at 12,000 × g for 20 min at 4°C. The supernatant was harvested, and the protein concentration was determined using the BCA kit (Beyotime). SDS sample buffer containing β-mercaptoethanol was added to the protein lyses. Totally 30 µg protein was loaded per lane, resolved by 8–12% polyacrylamide gels in a Mini-PROTEAN^® Tetra Vertical Electrophoresis Cell (Bio-Rad Laboratories, Inc., Hercules, CA, USA), electroblotted onto polyvinylidene fluoride membranes (Merck Millipore, Darmstadt, Germany) in a Mini Trans-Blot^® Module (Bio-Rad Laboratories), blocked with 5% skimmed milk/TBST, and incubated with the primary antibodies diluted in the blocking buffer overnight at 4°C. The following primary antibodies were used: mouse monoclonal anti-human hEPO (catalog no. SAB5300475; Sigma-Aldrich; Merck KGaA), rabbit polyclonal anti-human/mouse/rat p-EPOR (catalog no. SAB4301492; Sigma-Aldrich; Merck KGaA), rabbit polyclonal anti-human/mouse/rat p-ERK1/2 (catalog no. E7028; Sigma-Aldrich; Merck KGaA), rabbit polyclonal anti-human/mouse/rat p-AKT (catalog no. SAB4301497; Sigma-Aldrich; Merck KGaA), rabbit polyclonal anti-human/mouse/rat Nrf2 (catalog no. SAB4501984; Sigma-Aldrich; Merck KGaA), rabbit polyclonal anti-human/mouse/rat KEAP1 (catalog no. AV34727; Sigma-Aldrich; Merck KGaA), rabbit polyclonal anti-human/mouse/rat NQO1 (catalog no. N5288; Sigma-Aldrich; Merck KGaA), rabbit polyclonal anti-human/mouse/rat p62 (catalog no. P0067; Sigma-Aldrich; Merck KGaA), rabbit polyclonal anti-human/mouse/rat LC3B (catalog no. 3868; Cell Signaling Technology, Inc.) and rabbit polyclonal anti-human/mouse/rat β-actin (catalog no. SAB2100037; Sigma-Aldrich; Merck KGaA) (IgGs; 1:1,000). β-actin expression was used as a loading control. After the washout of the unbound primary antibodies, the membranes underwent incubation with the secondary antibodies [horseradish peroxidase (HRP)-conjugated goat polyclonal anti-mouse IgG (1:5,000; catalog no. 31430; Invitrogen; Thermo Fisher Scientific, Inc.) or HRP-conjugated goat polyclonal anti-rabbit IgG (1:5,000; catalog no. 31460; Invitrogen; Thermo Fisher Scientific, Inc.)] for 1 h at room temperature. Protein blots were visualized using a ChemiDoc™ Imaging System (Bio-Rad), and quantified using Image-Pro Plus 6.0.

Data are expressed as the mean ± standard error of the mean (SEM). Statistical analyses were performed using GraphPad Prism 7.0 software (GraphPad Software, Inc., La Jolla, CA, USA). Statistical differences were determined by one-way analysis of variance (ANOVA) followed by Bonferroni's multiple comparisons tests (more than two experimental conditions) and Wilcoxon signed-rank test (when compared with 50% chance level). P<0.05 was considered statistically significant.

To evaluate the impacts of rhEPO administered via intranasal route on the motor function of mice exposed to alcohol for 4 weeks, rotarod and beam walk tests were conducted after a 4-week rhEPO treatment. In the rotarod test (Fig. 1A), mice exposed to sustained alcohol in the CAE group showed dramatically shortened time remaining on the rotarod when compared with mice in the Control group (CAE vs. Control, P<0.001), whereas alcohol-exposed mice receiving intranasal rhEPO administration in the CAE+rhEPO group presented with significantly longer time maintaining on the rotarod (CAE+rhEPO vs. CAE, P<0.05), up to the level of the Control group (CAE+rhEPO vs. Control, P>0.05). When challenged on a 1-cm beam, mice in the CAE group needed more time to cross the beam than those in the Control group (CAE vs. Control, P<0.001; Fig. 1B). In contrast, mice in the CAE+rhEPO group spent less time crossing the beam (CAE+rhEPO vs. CAE, P<0.05) and regained similar beam-crossing capacity to that of mice in the Control group (CAE+rhEPO vs. Control, P>0.05). Data for the motor tests indicated that intranasal rhEPO treatment could promote recovery from chronic alcoholism-related motor deficits.

To examine the effects of intranasal rhEPO on cognition and anxiety following chronic alcoholism, mice in each group were tested in object location and open-field tasks post rhEPO treatment (4 weeks post alcohol exposure). During the acquisition phase in the novel location recognition task (Fig. 1C), mice in all groups spent the same amount of time exploring the objects. During memory testing 24 h later (Fig. 1D), mice in the Control and CAE+rhEPO groups similarly explored preferentially the object that had been moved to a novel location (CAE+rhEPO vs. Control, P>0.05; Control and CAE+rhEPO vs. chance level 50%, P<0.01), but mice in the CAE group showed no exploratory preference for the displaced object (CAE vs. Control, P<0.001; CAE vs. CAE+rhEPO, P<0.01; CAE vs. chance level 50%, P>0.05), indicating that the impaired discriminative memory of mice with chronic alcoholism for the novel and familiar location could be rescued by intranasal rhEPO. During the open-field test (Fig. 1E and F), alcohol-exposed mice without drug treatment had sharply diminished forelimb rears (CAE vs. Control, P<0.001), decreased time spent in the center and consequentially increased time spent in the corners of the open field (CAE vs. Control, P<0.001) compared to the healthy control mice, indicative of reduced exploration and aggravated anxiety. However, intranasal rhEPO administration remarkably restored vertical exploration activity (CAE+rhEPO vs. CAE, P<0.001; CAE+rhEPO vs. Control, P>0.05), increased center exploration and relieved anxiety (CAE+rhEPO vs. CAE, P<0.001; CAE+rhEPO vs. Control, P>0.05) in mice exposed to alcohol for 4 weeks. These findings in cognitive and behavioral tests suggested that intranasal rhEPO treatment could ameliorate the memory, learning and anxiety disorders in mice with chronic alcoholism.

To further determine the effects of intranasal rhEPO on nervous conduction after chronic alcohol exposure, somatosensory-evoked potential (SSEP) analysis was performed. In this study, the amplitude and latency of SSEP were selected as the objective measurements of nervous conduction. At week 4 after alcohol exposure (immediately after rhEPO treatment), mice in all groups exhibited evident SSEP waves in response to forepaw stimulation (Fig. 1I). Furthermore, alcohol-exposed mice with no drug treatment in the CAE group showed an obvious decrease in amplitude of forepaw-evoked potential (0.85±0.04 mV) when compared with those in the Control group (Control: 2.52±0.14 mV; CAE vs. Control, P<0.001; Fig. 1G and I), indicative of impaired ascending conduction caused by chronic alcoholism. Whereas intranasal EPO-treated mice presented with notably recovered SSEP amplitude (CAE+rhEPO: 2.24±0.11 mV; CAE+rhEPO vs. CAE, P<0.001) which became comparable to that of control mice (CAE+rhEPO vs. Control, P>0.05). Coinciding with the change in amplitude, the SSEP latency of alcohol-exposed non-treated mice (14.00±0.45 msec) was significantly longer than that of non-alcohol-exposed mice (Control: 8.46±0.44 msec; CAE vs. Control, P<0.001; Fig. 1H and I), indicating the delayed conduction of nervous impulse. Inconsistent with the results of amplitude analysis, although the conduction velocity of forepaw-evoked action potential was restored following intranasal EPO treatment (CAE+rhEPO: 10.36±0.50 msec; CAE+rhEPO vs. CAE, P<0.001), the SSEP conduction in EPO-treated mice was still slower than that of control mice (CAE+rhEPO vs. Control, P<0.05). The results of electrophysiological examination verified the therapeutic effects of intranasally administered EPO on the nervous conduction after chronic alcoholism.

Intranasal rhEPO promotes the remyelination in mice with chronic alcoholism. Continuous demyelination and failed remyelination are involved in the prolonged motor coordination impairments and progressive cognitive decline after brain injury (26). As revealed by the immunostaining for MBP (a major constituent of the myelin sheath) performed 4 weeks after alcohol exposure (just after the 4-week rhEPO treatment), the deep layers (V/VI) of the cerebral cortex of mice in the Control group were highly myelinated (Fig. 2A), consistent with a previous description (27). Alcohol-exposed mice in the CAE group showed comparatively impaired myelination in the same cortical layers, which was demonstrated by the decrease in the MBP⁺ myelin sheath (CAE vs. Control, P<0.001; Fig. 2A and B). When the alcohol-exposed mice were subjected to intranasal EPO treatment, the cortical MBP⁺ myelin level was significantly recovered but did not attain the normal level (CAE+rhEPO vs. CAE, P<0.001; CAE+rhEPO vs. Control, P<0.05). In the CA3 area of the hippocampus rich in myelinated nervous fibers (28), the deteriorated myelination occurred in mice that had been exposed to sustained alcohol (CAE vs. Control, P<0.001), which was approximately completely reversed in the intranasal EPO-treated mice as determined by stereological volume analysis of MBP fluorescence intensity (CAE+rhEPO vs. CAE, P<0.001; CAE+rhEPO vs. Control, P>0.05; Fig. 2C and D). The data for the immunofluorescence staining for the structural protein of myelin sheath demonstrated the positive effects of intranasal EPO on remyelination in chronic alcoholism-damaged brain.

Region-specific synapse loss is closely correlated to interrupted transmission of nervous pulse and impaired cognition in brain injury and neurodegeneration (29–31). To examine the cortical synapse formation which determines cortical synaptic transmission and consequently motor and cognitive function, the immunostaining for synaptophysin (a component of synaptic vesicle) was performed 4 weeks post alcohol exposure (12 h post the end of rhEPO treatment). Quantification of synaptophysin puncta revealed substantial synapse loss in the cortex of the chronic alcohol-exposed mice as compared to their healthy counterpart controls (CAE vs. Control, P<0.001; Fig. 3A and B). After intranasal administration of EPO, the chronic alcohol exposure-affected mice showed significantly elevated fluorescence intensity of synaptophysin in the cortex which was almost equivalent to that of the control mice (CAE+rhEPO vs. CAE, P<0.001; CAE+rhEPO vs. Control, P>0.05), suggestive of reconstructed cortical synapse formation supporting the exact nervous conduction and sophisticated neural functions. To further evaluate the cognition-related hippocampal synaptogenesis after alcohol exposure and EPO treatment, TEM analysis was conducted to recognize the ultramicroscopic structures of synapse. As determined by presynaptic (synaptic vesicle) and postsynaptic compartments (postsynaptic density) in electron micrographs, hippocampal synapse density obviously declined on prolonged alcohol exposure in mice of the CAE group in contrast to that in controls (CAE vs. Control, P<0.001; Fig. 3C and D), but synapse amount recovered on intranasal EPO treatment and reached the non-alcohol-affected level in the CAE+rhEPO group (CAE+rhEPO vs. CAE, P<0.001; CAE+rhEPO vs. Control, P>0.05), which was consistent with the tendency obverse in cortex. These data clarified the repairing potential of intranasal EPO for brain synaptogenesis disrupted in chronic alcoholism.

Intranasal rhEPO exerts neuroprotective effects on mice with chronic alcoholism by regulating the autophagy-related degradation of Nrf2. To establish the mechanisms underlying the neuroprotective effects of intranasal rhEPO, western blot analyses were performed on the microdissected cortex and hippocampus 30 min after EPO administration for the last time. As expected, hEPO was detected in the brain tissue of intranasal EPO-treated mice rather than non-treated mice (Fig. 4A). Correspondingly, the exogenous EPO caused a considerable increase in phosphorylated (p)-EPOR, p-ERK1/2 and p-AKT levels in the cerebrum of the treated mice (CAE+rhEPO vs. CAE, all P-values <0.001; CAE+rhEPO vs. Control, P-values <0.001; Fig. 4A-D), reminiscent of the augmented activation of EPOR, ERKs and PI3K/AKT signaling. Furthermore, intranasally delivered EPO led to the upregulated level of cerebral antioxidant and neuroprotective Nrf2, in parallel with the reduced level of Nrf2 inhibitor protein KEAP1 and the elevated protein expression of Nrf2 target gene NQO1 (CAE+rhEPO vs. CAE, all P-values <0.001; Fig. 4A and E-G), which suggested the increased Nrf2 activity by EPO. However, mere chronic alcohol exposure resulted in a detectable reduction in basal EPOR, ERK1/2, AKT and Nrf2 activity (CAE vs. Control, all P-values <0.05). As autophagy-degraded p62 is reported to enhance Nrf2 signaling by inactivating KEAP1 (15,17), the relationship between Nrf2 activity and autophagy progress was monitored. The accelerated degradation of p62 by autophagy and the autophagy-indicative accumulation of LC3-II in alcohol-affected brain tissue (CAE vs. Control, both P-values <0.05; Fig. 4A, H and I) was reversed by intranasal rhEPO (CAE+rhEPO vs. CAE, all P-values <0.001), coinciding with the EPO-activated autophagy-inhibitory ERKs and PI3K/AKT signaling, the reduced level of Nrf2-inhibiting KEAP1, and the rescued expression of Nrf2 and Nrf2-targeted NQO1. These observations indicated the involvement of antioxidant Nrf2 negatively regulated by the alcohol-induced autophagic response in the protective impacts of EPO.

To further determine whether EPO functions as neuroprotector by modulating Nrf2 activity, ATRA, the specific antagonist of Nrf2 (16,32), was administered intraperitoneally 30 min before intranasal EPO treatment for each time. The effective suppression of Nrf2 activity by ATRA was demonstrated by the downregulated protein level of NQO1 (CAE+rhEPO+ATRA vs. CAE+rhEPO, both P-values <0.001; Fig. 4A and G). Expectedly, the levels of intracerebral hEPO, activated EPOR and regulatory molecules (ERKs and AKT), autophagy-related proteins (p62 and LC3-II), KEAP1 and also Nrf2 were not influenced by ATRA-induced Nrf2 inhibition (CAE+rhEPO+ATRA vs. CAE+rhEPO, all P-values >0.05; Fig. 4A-C and F-H). However, if pre-treated by ATRA, the intranasal EPO-administered mice with chronic alcohol exposure showed significantly curtailed duration maintaining on the rotarod (Fig. 1A), delayed beam-crossing (Fig. 1B), abolished preference for displaced object in novel location recognition task (Fig. 1D), impaired exploration-learning activity and increased anxiety in open-field test (Fig. 1E and F), retarded conduction of SSEP with amplitude loss (Fig. 1G-I), decreased MBP⁺ myelin sheath in cerebral cortex (Fig. 2A and B) and hippocampus (Fig. 2C and D), diminished cortical synaptophysin level (Fig. 3A and B) and reduced hippocampal synapse density (Fig. 3C and D) (CAE+rhEPO+ATRA vs. CAE+rhEPO, all P-values <0.05), which suggested that the EPO-induced recovery in motor cooperation, cognition, nervous conduction and cerebral myelination and synaptogenesis were almost abrogated by the inhibition of Nrf2 activity. These data confirmed that the therapeutic effects of intranasal EPO on chronic alcoholism-related brain disorders were mediated, at least partially, by impeding the autophagy-related degradation of Nrf2.