Proteomic analysis of erythropoietin-induced changes in neuron-like SH-SY5Y cells Eritropoetin'in Nöron Benzeri SH-SY5Y Hücrelerinde İndüklediği Değişikliklerin Proteomik Analizi

Objective: Erythropoietin (EPO) is widely used for treatment of anemia associated with different diseases; however, its adverse effects limit its use in clinical prac-tice. Therefore, understanding the effects of EPO at the molecular and cellular level is crucial to adjust treatment regimes, and to develop non-hematopoietic EPO derivatives. In this study, we used a proteomics approach to identify how EPO treatment modifies the cellular proteome. Methods: SH-SY5Y neuroblastoma cells were used as the model system to analyze the effects of EPO treatment at different time points (24 h and 48 h). Proteomic analysis revealed changes in 74 proteins after EPO treatment. Following proteomics analysis, Reactome pathway analysis were carried out to identify the affected cellular pathways. Results: According to results, EPO alters the levels of 74 protein species (40 were increased, 34 were decreased). The levels of 35 proteins were changed by 24 h EPO incubation, whereas 17 protein species were altered by 48 h EPO incubation. Levels of 22 protein species were altered by both of the incubation periods (24 h and 48 h). Conclusion: Overall, our results suggest that EPO mainly affects protein species in glucose metabolism, protein and RNA metabolism, cytoskeletal proteins, and mitochondrial protein species. türlerini, hücre iskeleti proteinlerini ve mitokondrial protein türlerini etkilediği gösterilmiştir. Anahtar Kelimeler: Eritropoetin; Nöron; Proteomikler; Nöroproteksiyon;


Introduction
Erythropoietin (EPO) is a multifunctional cytokine, which is primarily produced by peritubular cells in the kidney [1]. EPO production is subject to transcriptional regulation, where hypoxia-inducible transcription factors stimulate EPO gene expression [2]. The primary function of EPO is to regulate production of erythrocytes in the bone marrow. EPO receptor is expressed in various tissues [3,4], including the nervous system, and EPO plays important roles in neurodevelopment [5,6] and neuroprotection [7][8][9].
Recombinant EPO is in clinical use for treatment of anemia associated with different diseases, including but not limited to AIDS [10,11], cancer [12,13], and renal failure [14,15]. However, long-term EPO administration is associated with several adverse events, including hypertension [16], cerebral convulsion, hypertensive encephalopathy, thrombo-embolism, iron deficiency, and influenza-like syndrome [17]. Given these findings, it is not only necessary to determine the optimal dose and time for EPO administration, but also to understand how EPO exerts its effects at the molecular level.
Previous studies on EPO has also benefited from proteomics to identify the interactions between cellular proteins and EPO [18], and to determine differentially regulated protein spots after EPO treatment or withdrawal [19][20][21][22][23]. However, the question of how EPO alters the neuronal proteome has not been addressed so far. Therefore, our aim was to use a proteomics approach to identify how EPO modifies the cellular proteome. We used SH-SY5Y neuroblastoma cell line as the in vitro model system to analyze the effects of EPO. Our findings show that EPO treatment alters the cellular proteome, mainly through the proteins related to glucose metabolism, mRNA metabolism, and protein metabolism [24,25].

EPO preconditioning
Twenty-four hour prior to treatment, 3 × 10 6 SH-SY5Y cells (Passage #15) were seeded in 75 cm 2 cell culture flasks in triplicate. Cells were treated with EPO (1 U/mL final concentration) for 24 h, 48 h or left untreated. Cells were maintained in reduced serum medium (OptiMEM, Gibco, Gaithersburg, MD, USA) during EPO treatment.

Sample preparation for 2DE electrophoresis
At the end of incubation time, sample preparation was performed according to Proteome Factory's 2DE sample preparation protocol for cell culture. Total protein was isolated from plated SH-SY5Y cells using TriPure Isolation Reagent (Roche Diagnostics, Mannheim, Germany) according to the manufacturer's protocol. Urea, ampholytes and DTT were added to a final concentration of 9 M urea, 2% ampholytes and 70 mM DTT during thawing of the protein-pellet (200 μg). After incubation for 30 min at room temperature and centrifugation for 45 min at 15,000 × g the supernatant was removed and frozen in new tubes at -80°C.

Gel image analysis
The 2DE gels used for comparison analysis were digitized at a resolution of 150 dpi using a PowerLook 2100XL scanner with transparency adapter. Two-dimensional image analysis was performed using the Proteomweaver software (Definiens AG, Munich, Germany). Protein spots with different spot intensities were selected according to two parameters: -a minimal significant factor which was evaluated as follow with a replicate quality test: "Based on the 500 highest intensity spot-pairs, an average replicate deviation of 68.25% was found. The standard deviation of the average intensities for a group with three gels is 23.67%. The regulation factor between two such groups has a standard deviation of 35.04%. The selected confidence level (0.05) results in a trust factor of 1.96. Exponentiating the standard deviation of the regulation factors with the trust factor results in a minimal significant regulation factor of: 1.802. A minimal significant factor of 1.802 was therefore applied for the selection of changed protein spots.

Trypsin-digestion/nanoLC-ESI-MS/MS
Protein identification using nanoLC-ESI-MS/MS was performed by Proteome Factory (Proteome Factory AG, Berlin, Germany). The HPLC system was coupled to MS detection via Qstar XL mass spectrometer (ABI, Foster City, CA, USA). Peptides from enzymatic cleavage were acified with formic acid and applied to nanoLC-ESI-MS/MS. After trapping and desalting the peptides on enrichment column (Zorbax SB C18, 0.3 × 5 mm, Agilent) using 1% acetonitrile/0.5% formic acid solution for 5 min peptides were separated on Zorbax 300 SB C18, 75 μm × 150 mm column (Agilent, Waldbronn) using an acetonitrile/0.1% formic acid gradient from 5% to 40% acetonitrile within 40 min. MS overview spectra were automatically taken with a tolerance of ± 50 ppm according to manufacturer's instrument settings for nanoLC-ESI-MS/MS analyses, peptide fragmentation and detection was accomplished with an accuracy of ± 0.5 Da. Proteins were identified using MS/MS ion search of the Mascot search engine (Matrix Science, London, England) and nr protein database (National Center for Biotechnology Information, Bethesda, MD, USA). Ion charge in search parameters for ions from ESI-MS/MS data acquisition were set to "1 +, 2 + or 3 +" according to the instrument's and method's common charge state distribution.

Pathway analysis
Reactome pathway database (access: http://www.reactome.org/) was used to identify which cellular pathways are affected by EPO treatment. Two time points (24 h and 48 h) were individually analyzed.

Statistical analysis
Mann-Whitney U-test has been used to compare the differences in spot intensities. p-Values < 0.05 were considered to be statistically significant.

Differentially regulated protein spots after EPO treatment
To identify how EPO affects global protein levels in SH-SY5Y cells, we treated the cells with EPO (1 U/mL) for 24 h and 48 h. This dose is consistent with previous studies, which characterized the effects of EPO in vitro [27,28]. Proteomic analysis revealed a total of 74 differentially speciated proteins after EPO treatment (Figures 1-3). The encircled spots are normally the regulated or the analyzed spots. The pH range is from 3 to 10 and the molecular weight is from 10 to 150 kDa. Among all differentially speciated proteins, 40 proteins were upregulated, and 34 proteins were downregulated. Thirty-five proteins showed differential speciation after 24 h EPO treatment (Tables 1 and 2), whereas only 17 proteins showed differential expression after 48 h EPO treatment (Tables 3 and  4). Twenty-two proteins showed differential speciation in both 24 h and 48 h EPO treatment (Tables 5 and 6). At 24 h time point, neuroleukin was the most upregulated protein (3.19-fold compared to control), whereas pyruvate kinase (PK) was the most downregulated protein (0.19-fold compared to control). At 48 h time point, heterogeneous nuclear ribonucleoprotein R isoform 2 was the most upregulated protein (2.35-fold compared to control), whereas peroxiredoxin-4 was the most downregulated protein (0.32-fold compared to control). The complete list of differentially speciated proteins is shown in SI Figures.

EPO affects proteins in different cellular pathways
In the next step, we used Reactome pathway database to identify the cellular pathways which are most affected by EPO treatment. The results indicate that EPO treatment affects proteins in glucose metabolism, protein and mRNA metabolism, cytoskeletal and mitochondrial proteins (Figures 4 and 5).

Discussion
Following its introduction to the clinical practice, the use of recombinant human EPO has increased exponentially in the past 20 years. EPO treatment is currently used for a broad spectrum of diseases. EPO has a direct effect on blood pressure [29], which is independent of its hematopoietic effects [16]. In addition, several in vitro, in vivo, and clinical studies have been carried out to characterize the hypertensinogenic effects of EPO [30][31][32][33].
The choice of SHSY-5Y cells was based on previous studies, which have demonstrated the expression of EPO receptor and tissue-protective EPO receptors in this cell line [34,35]. In addition, several studies have used SHSY-5Y cells to analyze the functional effects of EPO [36,37]. Therefore, we preferred using this cell line as a model to characterize the effects of EPO on cellular proteome.
Among all differentially speciated proteins after EPO treatment, PK showed the highest level of downregulation (approximately 80%). PK is one of the key metabolic enzymes that regulate glycolysis and gluconeogenesis, and its deficiency is the second leading cause of enzyme-deficient hemolytic anemia [38,39]. So far, there is no direct experimental evidence on the interplay between EPO and PK deficiency. Only a single study has investigated the  potential link between EPO and PK, which indicates that EPO is an effective treatment option for iron overload [40], a common event in PK deficiency [41,42].
Stathmin is a small, cytoplasmic protein that is primarily responsible for regulating the dynamics of microtubule formation. Stathmin functions as a relay protein in different signaling pathways (mainly PI3K and MAPK signaling) [43]. The functional roles of stathmin in hematopoiesis have been characterized in several studies. Stathmin speciation increases during chemically induced and EPO-induced erythroid differentiation. In this study, Glucose-6-phosphate isomerase (GPI) encodes an enzyme, which functions as a neurotrophic factor (known as neuroleukin) outside the cell. Neuroleukin is essential for survival of skeletal motor neurons and sensory neurons. In the present study, we have shown that the highest upregulation following 24 h EPO treatment was observed in neuroleukin levels. Thus, it is possible that EPO exerts its neuroprotective effects via neuroleukin.
Sideroflexin-3 is a member of the sideroflexin protein family. Sideroflexin-3 is located on the mitochondria, and functions as an iron transporter to regulate iron homeostasis. In our study, we determined that sideroflexin-3 speciation was the most upregulated protein after EPO treatment at both time points (24 h and 48 h). Given its role in iron homeostasis, sideroflexin-3 may function as a mediator of EPO's hematopoietic effects.  Identified protein species from SH-SY5Y samples, 24 h treatment of EPO with upregulated speciation calculated from improved spot density as percentage changes in spot volume against untreated SH-SY5Y samples.      Reactome pathway database identified the cellular pathways which were at most affected by EPO treatment. The results indicated that EPO treatment affected protein species in glucose metabolism, protein and mRNA metabolism as well as cytoskeletal and mitochondrial protein species. Each pathway was associated with annotated protein species through colored blocks and Reactome's pathway hierarchy was shown using statistics.

Concluding remarks
Taken together, our findings indicate that EPO affects more than 70 proteins, which play a role in multiple cellular mechanisms in neuronal cells. Functional characterization of these proteins is necessary to further elucidate the mechanism of action of EPO.