Global Proteomics Analysis of Bone Marrow: Establishing Talin-1 and Centrosomal Protein of 55 kDa as Potential Molecular Signatures for Myelodysplastic Syndromes

Myelodysplastic syndrome (MDS) is a hematological disorder characterized by abnormal stem cell differentiation and a high risk of acute myeloid leukemia transformation. Treatment options for MDS are still limited, making the identification of molecular signatures for MDS progression a vital task. Thus, we evaluated the proteome of bone marrow plasma from patients (n = 28) diagnosed with MDS with ring sideroblasts (MDS-RS) and MDS with blasts in the bone marrow (MDS-EB) using label-free mass spectrometry. This strategy allowed the identification of 1,194 proteins in the bone marrow plasma samples. Polyubiquitin-C (UBC), moesin (MSN), and Talin-1 (TLN1) showed the highest abundances in MDS-EB, and centrosomal protein of 55 kDa (CEP55) showed the highest relative abundance in the bone marrow plasma of MDS-RS patients. In a follow-up, in the second phase of the study, expressions of UBC, MSN, TLN1, and CEP55 genes were evaluated in bone marrow mononuclear cells from 45 patients by using qPCR. This second cohort included only seven patients from the first study. CEP55, MSN, and UBC expressions were similar in mononuclear cells from MDS-RS and MDS-EB individuals. However, TLN1 gene expression was greater in mononuclear cells from MDS-RS (p = 0.049) as compared to MDS-EB patients. Irrespective of the MDS subtype, CEP55 expression was higher (p = 0.045) in MDS patients with abnormal karyotypes, while MSN, UBC, and TALIN1 transcripts were similar in MDS with normal vs. abnormal karyotypes. In conclusion, proteomic and gene expression approaches brought evidence of altered TLN1 and CEP55 expressions in cellular and non-cellular bone marrow compartments of patients with low-risk (MDS-RS) and high-risk (MDS-EB) MDSs and with normal vs. abnormal karyotypes. As MDS is characterized by disrupted apoptosis and chromosomal alterations, leading to mitotic slippage, TLN1 and CEP55 represent potential markers for MDS prognosis and/or targeted therapy.

Myelodysplastic syndrome (MDS) is a hematological disorder characterized by abnormal stem cell differentiation and a high risk of acute myeloid leukemia transformation. Treatment options for MDS are still limited, making the identification of molecular signatures for MDS progression a vital task. Thus, we evaluated the proteome of bone marrow plasma from patients (n = 28) diagnosed with MDS with ring sideroblasts (MDS-RS) and MDS with blasts in the bone marrow (MDS-EB) using label-free mass spectrometry. This strategy allowed the identification of 1,194 proteins in the bone marrow plasma samples. Polyubiquitin-C (UBC), moesin (MSN), and Talin-1 (TLN1) showed the highest abundances in MDS-EB, and centrosomal protein of 55 kDa (CEP55) showed the highest relative abundance in the bone marrow plasma of MDS-RS patients. In a follow-up, in the second phase of the study, expressions of UBC, MSN, TLN1, and CEP55 genes were evaluated in bone marrow mononuclear cells from 45 patients by using qPCR. This second cohort included only seven patients from the first study. CEP55, MSN, and UBC expressions were similar in mononuclear cells from MDS-RS and MDS-EB individuals. However, TLN1 gene expression was greater in mononuclear cells from MDS-RS (p = 0.049) as compared to MDS-EB patients. Irrespective of the MDS subtype, CEP55 expression was higher (p = 0.045) in MDS patients with abnormal karyotypes, while MSN, UBC, and TALIN1 transcripts were similar in MDS with normal vs. abnormal karyotypes. In conclusion, proteomic and gene

INTRODUCTION
Myelodysplastic syndromes (MDSs) are hematologic stem cell malignancies associated with cytopenias, bone marrow insufficiency, and anemia (1). The pathogenesis of MDS relates to disruption of stem cell development, increased apoptosis, mutations in splicing factors and DNA repair system, altered DNA methylation of tumor suppressor genes or protooncogenes, and immune derangement (2)(3)(4). According to the Revised International Prognostic Scoring System (IPSS-R), MDS is grouped into very low-, low-, intermediate-, very high-, and high-risk disorders, depending on the pathogenesis and risk of acute myeloid leukemia (AML) transformation (5).
MDS with ring sideroblasts (MDS-RS) is characterized by bone marrow dysplasia, cytopenia, low risk of AML transformation, ring sideroblasts ≥15% of bone marrow precursors, and presence of SF3B1 mutation (6,7). Ring sideroblasts are erythroid precursors with abnormal accumulation of iron inside mitochondria, and MDS-RS cases have stem cells with abnormal erythroid colony formation with failure of terminal erythroid differentiation (TED) (8). TED was also reported in high-risk MDS (8), characterized by excess blasts (MDS-EB) and 5%-19% myeloblasts in the bone marrow and accounts for up to 40% of all MDS cases, with a high risk of AML transformation (6,9,10). Low-risk MDS, such as MDS-RS, commonly presents symptoms of anemia and proper response to erythropoietin treatment (8). MDS-EB is linked to high-risk karyotypes such as monosomy seven, complex karyotype and mutations of genes that control disturbance of apoptosis (TP53) and DNA repair, stem cell maturation (RUNX1), and histone modifications (ASXL1) (11).
The bone marrow microenvironment contains cellular and non-cellular compartments that modulate the functional status of hematopoietic stem cells (12), and soluble factors present in the bone marrow modulate cell-cell and cell-extracellular matrix interactions and intracellular signaling (13). Alterations in their biochemistry and synthesis of bone marrow factors interfere with hematopoiesis and potentially trigger aberrant mechanisms that define the pathobiology of MDS. In fact, a recent study reports that pronounced differences exist in the proteomes of bone marrow plasma of patients with AML and healthy ones (14), with upregulated and downregulated proteins associated with chemokine and cytokine signaling pathways. Protein profiles are also altered in the blood and bone marrow plasma of patients with low-and high-risk lymphocytic leukemia (Braoudaki et al. (15)). The proteome of bone marrow cellular compartments and blood plasma are also altered in patients with AML (16), and plasma from patients diagnosed with MDS has decreased levels of CXC chemokine ligands 4 and 7 (17). Analyses of plasma samples using 2-D sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and nano-liquid chromatographytandem mass spectrometry (nano-LC-MS/MS) identified 39 proteins with different abundances in MDS patients with refractory anemia and refractory anemia with ringed sideroblasts (18). The same research team reported 27 different proteins in the plasma of MDS patients with refractory anemia with excess blasts-2 in comparison to healthy individuals (19). Other studies used gel-based proteomics to describe protein profiles of platelets (20) and peripheral mononuclear cells (21) of patients with MDS. These results clearly indicate that hematological stem cell malignancies are preceded by and/or cause dramatic alterations in the proteome of diverse biological entities, from blood to components of the bone marrow.
Researchers have made remarkable contributions to the understanding of the pathophysiology of MDSs (4,(22)(23)(24). Despite this unprecedented effort, the current therapy with hypomethylating agents is not effective for all cases of the disease, especially for high-risk MDS (22,25), and the identification of molecular markers of MDS progression and response to treatment may be crucial for effective therapeutic strategies in precision medicine. Thus, a series of studies were designed to decipher potential molecular signatures of MDS subtypes. The first study evaluated the proteome of bone marrow plasma from patients diagnosed with low-risk (MDS-RS) and high-risk MDS (MDS-EB). As centrosomal protein of 55 kDa (CEP55), moesin (MSN), talin-1 (TLN1), and ubiquitin C (UBC) showed the highest differential abundances in MDS-RS vs. MDS-EB cases, a second phase of the research was carried out to analyze the expression of those respective genes in mononuclear cells of a distinct cohort of MDS patients.

Research Design and Ethical Statement
The present research consisted of two independent studies ( Figure 1). In Study 1, label-free data-dependent acquisition (DDA) mass spectrometry was used to decipher the proteome of bone marrow plasma from patients with MDS-RS and MDS-EB (n = 28). Tools of bioinformatics and statistical methods were employed for in silico analysis of protein functional attributes and protein differential quantities in the samples. Based on such analyses, CEP55, MSN, TLN1, and UBC were proteins with the highest differential abundances in the bone marrow plasma of MDS-RS and MDS-EB patients. Thus, Study 2 was further carried out to examine the expression of those genes in bone marrow mononuclear cells from an additional cohort of MDS patients (n = 45).
This study was approved by the Ethics Committee of the Federal University of Ceara, UFC (#69320817.7.0000.5054), with written consent from each participant. Patients were diagnosed with MDS at the Drug Research and Development Center, School of Medicine, UFC, and corresponded to treatmentnaive patients.

Cytogenetics
G-banding karyotype of patients from Study 1 and Study 2 (see below) was performed as previously reported (26). Briefly, cultures were established in Roswell Park Memorial Institute (RPMI) 1640 medium (Gibco, Grand Island, NY, USA) containing 30% fetal calf serum. For the 24-h culture, colcemid was added at a 0.05 mg/ml concentration for the final 30 min of culture. After being harvested, cells were exposed to a hypotonic KCl solution (0.068 mol/L) and fixed with Carnoy's buffer (1:3 acetic acid/methanol). Slides were stained with Giemsa solution, and 20 metaphases were analyzed whenever possible. The karyotype was prepared using CytoVision Automated Karyotyping System (Applied Imaging, Grand Rapids, MI, USA) and described according to the International System for Human Cytogenomic Nomenclature 2016 (27).

Study 1: Quantitative Proteomic Analysis of Bone Marrow Plasma
The proteome of bone marrow plasma was evaluated in samples from 13 patients with MDS-RS and 15 patients diagnosed with MDS-EB, which included subcategories MDS-EB-1 (5%-9% blasts in the bone marrow) and MDS-EB-2 (10%-19% blasts in the bone marrow) ( Table 1). Diagnosis of the disease followed the WHO 2016 guidelines (1), and the risk stratification was based on the IPSS-R (5).

Sample Collection and Preparation, Trypsinization, and Desalting
Samples were obtained by sternal aspiration of bone marrow by a clinical hematologist at the same time as trephine biopsy at diagnosis. The bone marrow was collected in 4-ml Vacutainer ™ tubes with spray-coated K 2 EDTA (Thermo Fisher, Waltham, MA, USA). Immediately after collection, tubes were inverted several times and centrifuged at 700 ×g (15 min, 4°C). Then, the supernatant (plasma) was transferred to another tube, treated with a protease inhibitor mix (Sigma-Aldrich, St. Louis, MO, USA), and centrifuged at 10,000 ×g (30 min, 4°C). Afterward, the plasma was pipetted into a clean tube and stored at −80°C. An aliquot of this plasma was used for soluble protein quantification (28). Based on this quantification, both 1-D and 2-D gels (SDS-PAGE) were run for assessment of general protein patterns (Supplementary Figure 1). Gels were prepared and run as reported before (29), with modifications.

Label-Free Mass Spectrometry
Three micrograms of tryptic peptides from each sample was individually applied to a Dionex Ultimate 3000 liquid chromatographer (Thermo Scientific, USA) for reversed-phase nano-chromatography, as previously described (31). The peptides were injected into a 2 cm × 100 mm trap column containing C18, 5-mm particles (Dr. Maisch GmbH, Ammerbuch, Germany). Then, peptides were eluted from this column to another analytical one (32 cm × 75 mm) containing C18, 3-mm particles (Dr. Maisch GmbH, Germany), and finally eluted to the spectrometer's ionization source. The elution gradient was composed of 0.1% formic acid in water (solvent A) and 0.1% formic acid in acetonitrile (solvent B), in a gradient of 2% to 35% solvent B for 170 min.
Samples were analyzed in positive DDA mode in a label-free mass spectrometric approach using an Orbitrap Elite instrument (Thermo Fisher, USA) (31,32). The eluted fractions generated MS1 spectra between 300 and 1,650 m/z with a resolution of 120,000 FWHM at 400 m/z. The 20 most abundant ions from MS1 with at least two charges were automatically selected to fragmentation (MS2) by higher-energy collisional dissociation (HCD) with automatic gain control (AGC) of 1 × 10 (6) and

Data Analyses
MS1 spectra found in the chromatograms were aligned and quantified according to integrated intensity area from the extracted-ion chromatogram (XIC) peaks generated by the respective ion and quantified using Progenesis QI software Nonlinear Dynamics (Waters, Milford, MA, USA). Protein identification was performed using Peaks software, which deduces sequences from the fragmentation information and searches in the UniProt database. Protein identification information was inserted again in the Progenesis QI program and combined with quantitative data generated previously (31). Statistical analysis was performed using Progenesis QI software to evaluate differences in protein abundance between MDS-RS and MDS-EB samples. A first statistical analysis was performed before protein identification to filter the MS1 features presenting ANOVA p-values <0.05. Peaks 7.0 software was used with the fragmentation spectra, and the Homo sapiens UniProt database was searched. Search parameters were set as follows: precursor ion mass error tolerance of 10 ppm, MS/MS mass tolerance of 0.05 Da, carbamidomethylation of cysteine residues (fixed modification), deamidation, and methionine oxidation (variable modifications). A maximum of two missed cleavage sites per peptide was allowed. The identified proteins were filtered at a rate of 1% for false discovery rate (FDR), and a minimum of 1 unique peptide per protein was required for identification. Proteins with different abundances (p < 0.05) between MDS-RS and MDS-EB samples were subjected to multivariate analysis using MetaboAnalyst 4.0 (http://www. metaboanalyst.ca) (31,33,34). For this type of analysis, protein data were log transformed, normalized based on sample median, and mean centered. Principal component analysis and partial-least squares discriminant analysis (PLS-DA) with variable importance in projection (VIP) approaches were applied to verify data trends and outliers, protein discrimination, and importance in the MDS-RS vs. MDS-EB scenario, respectively.

Analysis of Functional Clusters
Functional clusters associated with bone marrow plasma proteins of patients diagnosed with MDS-RS and MDS-EB were analyzed through the DAVID platform (DAVID-Functional Annotation Bioinformatics Analysis-https://david. ncifcrf.gov) (35)(36)(37). For this analysis, UniProt accession numbers were uploaded to the DAVID platform, and clusters were defined according to enrichment scores and p-values [−log (p-value)].

Study 2: Analysis of Gene Expressions in Bone Marrow Mononuclear Cells
Based on the results of Study 1, UBC, MSN, and TLN1 proteins showed the highest abundances in MDS-EB, and CEP55 was identified with the highest comparative abundance in the bone marrow plasma of MDS-RS patients. As the next phase of the research, gene expression validation of CEP55, MSN, TLN1, and UBC was carried out by qPCR in bone marrow mononuclear cells from 45 MDS patients, including 15 MDS-RS and 23 MDS-EB patients, five with multilineage dysplasia, one with single lineage dysplasia, and one with MDS secondary to therapy ( Table 2). MDS patients were evaluated according to the IPSS-R (5). Four MDS-RS and three MDS-EB patients listed in Study 1 were included in Study 2.

Total RNA Extraction and Quantitative Real-Time PCR
Bone marrow was obtained by sternal aspiration of all patients, as described in Study 1. After centrifugation and removal of plasma, mononuclear cells were separated after lysis of red blood cells and subjected to total RNA extractions using Trizol Reagent ™ (Invitrogen, Carlsbad, CA, USA), followed by RNA quantification (NanoDrop 2000c; Thermo Scientific, USA). Complementary DNA (cDNA) was obtained using a High-Capacity cDNA Reverse Transcription kit (Applied Biosystems, Foster City, CA, USA) according to the manufacturer's recommendations. Quantitative real-time PCR (qPCR) reactions were performed using the TaqMan methodology (Applied Biosystems, USA) on 7500 Fast System ® (Applied Biosystems, USA). The custom TaqMan Gene Expression Assays were used for CEP55 (Hs01070181_m1), MSN (Hs00792607_mH), TLN1 (Hs00196775_m1), and UBC (Hs00824723_m1) and the reference gene B2M (beta-2microglobulin, Hs99999907_m1). B2M has been previously validated as a stable reference in gene expression analysis of bone marrow mononuclear cells from MDS patients (38). PCRs were performed using 1× TaqMan ® Universal Master Mix II, with UNG ® (Applied Biosystems, USA), 1× ml of each probe, and 2× ml of cDNA. The following cycling conditions were performed: initial denaturation at 95°C for 15 min followed by 40 cycles consisting of 95°C for 15 s and 60°C for 1 min. Each sample was performed in duplicate, and the expression ratios were calculated using the 2 −DCq method (39).

Statistical Analysis
Normality of mRNA data distribution was evaluated by the Shapiro-Wilk test and Student's t-test or one-way ANOVA with Tukey/Games-Howell post-hoc test used when normality was detected. The variance homogeneity for all variables analysis was evaluated by Levene's test. Statistical analyses were performed using SPSS 21.0 (SPSS Inc., Chicaog, IL, USA) and GraphPad Prism 8 (GraphPad Software, La Jolla, CA, USA) software.

In Silico Analyses of Protein-Protein
Interactions, MiRNA-Linked Networks, and CD34+ Gene Expression Networks of TLN1 and CEP55 proteins were analyzed using String platform version 11.5 (https://string-db.org/ (31,40)). Protein physical interactions, network type, network edges, and active interaction sources were based on String default settings,  (41). Protein physical interactions, co-expression, predicted networks, co-localization, pathways, genetic interactions, and shared protein domains were analyzed using the default GeneMANIA approach (automatically selected weighting method) and the Homo sapiens gene nomenclature. The potential regulation of TLN1 and CEP55 genes by miRNAs was analyzed using the miRNet 2.0 database (https:// www.mirnet.ca) (42). The network was built considering a miRNA database for humans and bone marrow as a specific tissue and using a 1.0 degree cutoff of a node. The degree represents the number of connections, and nodes with a higher degree are important hubs in a network. Moreover, functional enrichment of all miRNAs was built, with highlights for main biological functions.
We performed in silico evaluation of CEP55, TLN1, and MSN gene expression in CD34+ cell markers from MDS patients presented in the Gene Expression Omnibus (GEO) data bank (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc= GSE58831), as described by Gerstung    binding involved in cell-cell adhesion, actin binding, phosphatidylcholine binding, and conversion and serine-type endopeptidase activity, among others ( Figure 2; Supplementary Tables 1B, C).

Differentially Abundant Proteins in the Bone Marrow Plasma of Patients With Myelodysplastic Syndrome With Ring Sideroblasts and Myelodysplastic Syndrome With Excess Blasts
Based on the label-free mass spectrometry and data analysis by Progenesis QI software, there were 73 differentially abundant proteins (p < 0.05) in the bone marrow plasma of MDS-RS and MDS-EB patients (Supplementary Table 2). The partial least square discriminant analysis confirmed a pronounced contrast in the protein profiles of bone marrow plasma from MDS patients ( Figure 3A). According to VIP scores, proteins identified as Lactoferrin, Coagulation factor V, Polyubiquitin-C, Immunoglobulin heavy variable 3-66, Inositol 1,4,5trisphosphate receptor, MSN, Histone H2B type 1-J, cDNA FLJ56274, Kininogen 1 isoform, Talin-1, and Histone H1.5 made the most meaningful contributions to characterize the bone marrow plasma of MDS-EB patients. In contrast, Rheumatoid factor RF-IP12, Immunoglobulin kappa var. 1-6 IgG H chain, and CEP55 defined the major representation of bone marrow plasma from MDS-RS patients ( Figure 3B; Table 3). Considering fold-change values, Polyubiquitin-C (UBC), MSN, and Talin-1 (TLN1) showed the highest relative abundances in MDS-EB, and CEP55 was the major protein of the bone marrow plasma from MDS-RS patients ( Table 3;  Supplementary Table 2).

Protein Networks, MiRNA-Gene Interactions, and CD34+ Cell Gene Expression
As determined by in silico models, the most significant (score ≥ 0.977) components of TLN1 interactome included vinculin, actin filament (F-actin)-binding protein, integrin beta-1, 2 and 3, and   Figure 5A). In silico analysis using the GeneMANIA platform indicated that CEP55, MSN, TLN1, and UBC shared pathways, domains, and physical interactions ( Figure 5B). Based on the analysis conducted with the miRNet 2.0 database, TLN1 is regulated by onco-miRNAs and other miRNAs associated with cell cycle, bone regeneration, cell death, immune system, angiogenesis, and hematopoiesis, among others ( Figure 6). CEP55 gene was regulated by miRNAs associated with bone regeneration, DNA damage, cardiotoxicity, cell cycle, regulation of Akt pathway, and onco-miRNAs, among others ( Figure 6; Supplementary Table 4). In silico analysis of CD34+ cell gene expression showed that CPE55 expression was reduced while both TLN1 and MSN were increased in MDS patients compared to healthy controls (Figure 7; Supplementary Tables 1D, E).

DISCUSSION
Myelodysplastic syndromes are clinically heterogeneous and characterized by dysplasia of bone marrow cells, peripheral blood cytopenia, bone marrow insufficiency, disrupted hematopoiesis, and anemia, among other factors (1,(2)(3)(4). Molecular determinants of MDS pathobiology are complex and involve altered genomic and epigenetic factors, protein synthesis and structure, intracellular signaling, and mechanisms that control cell communication and fate. Therefore, our current results advance our knowledge about the molecular landscapes of MDS. Proteins were identified by label-free mass spectrometry and data, analyzed by Progenesis QI software. Listed proteins as depicted in Figure 3B.  Table 2.  Mass spectrometry coupled with nano-high-performance liquid chromatography (nano-HPLC) allowed the identification of 1,194 proteins in the bone marrow plasma of MDS patients. Gene ontology biological processes and molecular functions basically followed the same pattern for BM plasma proteins of MDS-RS and MDS-EB patients, including complement activation and immune-related events. In fact, chronic inflammatory diseases associated with activated innate immune signaling pathways usually precede MDS, creating the concept that this event in hematopoietic stem cells (HSC) relates to MDS etiology. Immune response-related genes in HSC are overexpressed in more than 50% of MDS patients (44), and our analysis demonstrates that response to stimulus and immune system process were major biological processes associated with the BM plasma proteome of patients with MDS-RS and MDS-EB. Functional clusters associated with the BM plasma proteome included, as expected, immune mechanisms, endocytosis, nucleosome assembly, chromatin silencing, blood coagulation and fibrinolysis, cytolysis, and cell-cell adhesion, among others. There is an important interplay between chronic immune stimulation and MDS pathogenesis. In agreement with this concept, a study conducted by Pellagatti et al. (45) reported upregulation of interferon-stimulated genes in CD34+ cells obtained from the bone marrow of MDS patients. There is also robust evidence of immune deregulation in MDS due to the presence of autoreactive T-cell clones and that interferon g had meaningful inhibitory effects on hematopoietic stem cells. We have previously detected, in a study with 111 MDS patients (46), that interferon regulatory genes are upregulated and linked to MDS pathogenesis, confirming the fundamental link between genomic instability and immune activation.
Considering the scenario with 1,194 proteins identified in the bone marrow plasma, univariate and multivariate statistical approaches grouped 15 proteins that most defined MDS-RS and MDS-EB subtypes. Among this select group, TLN1, MSN, and UBC had the greatest abundances in the bone marrow plasma of MDS-EB, and CEP55 showed the highest abundance in MDS-RS patients. CEP55 is a major player in cytokinesis, participates in PI3K/AKT signaling pathway, and is associated with tumor progression (47). MSN, in turn, connects actin filaments with the plasma membrane and contributes to cell shape, membrane transport, signal transduction, T-and B-cell homeostasis, and macrophage phagocytosis. Experimental evidence support UBC's roles in autophagy and signaling in cancer cells (48) and attachment of ubiquitin types to other proteins modulate DNA repair, cell-cycle communication, lysosomal and protein degradation, and kinase activation, among other events (49). As a cytoskeleton molecule, TLN1 connects integrin domains to actin and intermediate filaments, helping cell-cell and cell-extracellular matrix (ECM) attachment, cell migration, and signal transduction (50,51). In silico analysis based on the GeneMANIA platform showed pathways, domains, and physical interactions shared by CEP55, MSN, TLN1, and UBC, indicating that the actual work of these proteins certainly involves more complex mechanisms than those foreseeing when only their individual roles are analyzed.
Proteomics data gathered in Study 1 and the well-known bio tasks of CEP55, MNS, UBC, and TLN1 prompted us to investigate the expression of these genes in a follow-up study with 45 patients. In this case, TLN1 was overexpressed in mononuclear cells from MDS-RS when compared to MDS-EB patients. TLN1 contains alpha-helices and post-translational modification sites (https://alphafold.ebi.ac.uk/entry/Q9Y490), with actin-and integrin-binding domains (50,52,53). TLN1 is a player in mechanotransduction and filopodia function (54,55), and talin dysregulation results in pathological phenotypes with altered cell spreading, migration, and survival (56,57). Talin is associated with enhanced cell proliferation (common phenomena of blasts in MDS-EB), anoikis resistance and metastasis (57,58), hematologic disorders (54), and platelet and neutrophil activation (59). Talin-integrin interdependence affects endothelium linkage, migration of platelets and neutrophils (59,60) and leukocyte function (61), and TLN1 knockout mouse embryos have defective angiogenesis (62). From the perspective of clinical practice, we know that MDS-EB shows a high risk of AML transformation due to an increase in proliferation and survival of blasts. Thus, functional studies are needed to evaluate if TLN1 expression is related to AML transformation and, if so, this information will generate markers of MDS pathogenesis and possible targets of treatment.
Why, in our research, TLN1 was more abundant in the bone marrow plasma of MDS-EB and TLN1 gene was prevalent in mononuclear cells of MDS-RS patients is not clearly understood. Bone marrow plasma obviously contains molecules synthesized by BM cells and components from blood, suggesting that the content of BM plasma does not entirely reflect what comes from mononuclear cells. Moreover, mRNA and protein quantities in a given biological entity can be unrelated (63)(64)(65)(66), and cell lysis and/or apoptosis may have occurred during the collection and separation of BM plasma. Another possible explanation is that MDS-RS cases present serious problems of splice machinery (mutation of SF3B1), leading to disrupted mRNA translation, with increases in some mRNAs without proportional changes in protein synthesis. In this regard, comparisons between protein abundances in BM plasma and mRNA quantities in mononuclear cells are not straightforward because two levels of molecular information were actually generated in our studies, i.e., proteomics of BM plasma and qPCR in mononuclear cells. Thus, these technical approaches provide complementary and not conflicting pieces of information. High TLN1 abundance in the BM plasma of patients with a high-risk MDS subtype (MDS-EB) may relate to TLN1-mediated activation of FAK/AKT signaling and anoikis resistance (67). Low TLN1 gene expression detected in mononuclear cells of MDS-EB patients could be the result of impaired talin/integrin complex.
In silico analyses outline TLN1 interactions with proteins that regulate cell-cell and cell-ECM adhesion, cell migration, apoptosis, and intracellular signals, such as vinculin, integrins, and A4 precursor protein-binding family B member 1interacting protein (APBB1IP). Vinculin is downregulated in mononuclear cells of MDS patients (21), and APBB1IP is a RASrelated protein acting in high cell proliferation and AML transformation in MDS-EB cases (68). In silico models also identified that onco-miRNAs and miRNAs linked to cell cycle control and death, immune system, angiogenesis, and hematopoiesis are capable of regulating TLN1 expression. Some of those miRNAs, such as miR-17-3p and miR-17-5p, are upregulated in the BM of patients with MDS-RS, refractory cytopenia with multilineage dysplasia (RCMD), and RCMD with ringed sideroblasts, according to a recent review (69) and other studies (70)(71)(72). Other miRNAs, such as miR-27a-3p and miR-16-5p, were reduced in the plasma of patients with refractory cytopenia with unilineage dysplasia, refractory anemia with ringed sideroblasts, RCMD, MDS-EB1, and MDS-EB2. Hematopoiesis-related miR-34a-5p and miR-146b-5p were increased in plasma and extracellular vesicles of MDS patients with multilineage dysplasia, MDS-RS, MDS with isolated del(5q), MDS-EB 1, and MDS-EB 2. Given its functional properties and association with cancer processes (54,73), TLN1 is indeed a potential drug for targeted therapy (74), but more investigation is needed to decipher how drugs should be developed to tackle protein action and/or TLN1 expression. Considering that hematopoietic stem and progenitor cells (HSPCs) are important for MDS prognosis and treatment, we performed an in silico analysis of CD34+ gene expression and verified that CEP55 was reduced and both TLN1 and MSN expressions were increased in the bone marrow of MDS patients without any stratification.
As defined in Study 1, CEP55 was more abundant in the bone marrow plasma of MDS-RS patients. Remarkably, we observed in Study 2 that increased CEP55 expression in mononuclear cells was linked to chromosomal abnormality in MDS patients, regardless of MDS subtypes. CEP55 has domains (https:// alphafold.ebi.ac.uk/entry/Q53EZ4) involved in abscission, midbody localization (75), dimerization (76)(77)(78), and a ubiquitin-binding domain required for CEP55 action on cytokinesis (75,77). CEP55 influences aneuploid cells during perturbed mitosis (79), and high CEP55 expression in murine models facilitates exit from mitotic arrest, resulting in aneuploidy and resistance to anti-mitotic drugs. CEP55-depleted organisms have increased multinucleated cells and/or cells arrested at the midbody stage (80). Based on in silico modeling, CEP55 interacts with components of the ESCRT-I complex, a regulator of vesicular transport, cell growth, and differentiation; a microtubule-associated protein; a regulator of Rac GTPase activity and cell cycle during carcinogenesis; and cyclin-dependent kinases. This scenario reflects the multidimensional attributes of CEP55. Similar to TLN1, the expression of CEP55 is potentially regulated by oncogenes and miRNAs linked to DNA damage response, cell cycle, and AKT pathway. Several genes that define normal blood cells are regulated by miRNAs, and MDS prognosis is linked to certain miRNAs of the bone marrow, peripheral blood, and mononuclear cells (69). In fact, some CEP55-controlling miRNAs identified in our in silico model have been listed as differentially expressed in other patients with MDS (70,71,81,82).
Chromosomal abnormalities occur in up to 50% of MDS cases, mostly caused by problems in the spindle assembly checkpoint (SAC), which controls mitotic progression, chromosome alignment, and segregation. SACs delay anaphase until accurate kinetochore-microtubule attachment at metaphase and inhibition of SAC prolongs mitotic arrest. The majority of antimitotic chemotherapies alter microtubule dynamics, but cancer cells can bypass mitotic arrest and prematurely exit mitosis. High CEP55 expression is a marker of chromosomal alterations, aneuploidy, and poor prognosis of cancer patients (79,83,84) and promotes tumorigenesis in transgenic mice (85) as well. Clinical practice indicates that overall survival reaches 3 years in MDS patients with normal karyotypes, while cases with complex karyotypes show a 6-month survival. As CEP55 consistently relates to mitotic slippage in certain types of cancers, we urgently need to study the effects of CEP55 reduced expression and if CEP55 plays a central role in the genesis of cytogenetic abnormalities.
In conclusion, proteomic and gene expression approaches were used to unfold some of the molecular signatures related to MDSs. We recognize that our study has limitations, as we did not evaluate other MDS subtypes such as MDS with multilineage dysplasia, MDS secondary to therapy, and hypoplastic MDS. Also, gene expression was not analyzed in specific HSPC such as CD34+ cells, a marker for prognosis and treatment of MDS. To our knowledge, however, this is the first evidence of altered TLN1 and CEP55 in both cellular and non-cellular bone marrow compartments of patients with low-and high-risk MDS and with normal vs. abnormal karyotypes. The current diagnostic criteria and classification for MDS include bone marrow chromosomal alterations, blast count, and cytopenia, which are important to be analyzed for the prognosis of untreated patients. Given the complex and multifactorial characteristics of MDS, a deep molecular investigation of altered pathways provides a better understanding of the disease prognosis, its evolution to AML, and resistance to treatment. The novel potential markers CEP55 and TLN1 bring insights into the molecular signatures of MDS, setting the foundation for further investigations with stratified patients and HSPC and helping in the MDS prognosis and targeted therapy.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://massive.ucsd. edu/ProteoSAFe/static/massive.jsp MassIVE ID=MSV000088457.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Ethics Committee of the Federal University of Ceara, UFC (#69320817.7.0000.5054). The patients/participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

AUTHOR CONTRIBUTIONS
AAM, CLMF, AMAM, MOM-F, CP, and RFP designed the study and the overall organization of the article. RTGO and RFP collected the bone marrow samples and analyzed the karyotypes. MJBB, AMAM, WF, and RMO prepared the samples and carried out the proteomic analysis. MVS, WF, and CAOR provided the expertise and technical and financial support for the mass spectrometry, data search, and Progenesis analysis. DPB, RTGO, and CLMF did the RNA isolation from mononuclear cells and conducted qPCR and gene expression analysis. AGV, KMF, GGCC, and CRKP contributed to the bioinformatics analysis. MEAM, SMMM, MOM-F, AAM, CP, and RFP were responsible for the supervision of students and postdoctoral fellows and provided financial and institutional support, which was essential for the completion of the project. AAM, MJBB, CLMF, AMAM, CP, WF, RTGO and RFP contributed to the writing process. Final text editing was done by AAM, CLMF, AMAM, CP, and RFP. All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication.   (29)). Briefly, 13-cm strips (pH 3-10; GE Healthcare, USA) were rehydrated in a solution containing BM plasma proteins (300 µg) and subjected to isoelectric focusing. The second dimension was run as in (A). All 1-D and 2-D gels were stained with Coomassie blue and scanned at 300 dpi (Image Scanner, GE Healthcare, USA). Gel # 1 refers to BM sample from a patient with MDS with single lineage dysplasia (MDS-SLD); gels # 3, 6 and 7 refer to samples of BM plasma from MDS-RS patients; gels # 2, 4 and 5 contain BM plasma proteins from MDS-EB patients.

Supplementary Table 1 | (A)
Proteins of the bone marrow plasma from patients (as listed in ) with myelodysplastic syndrome. Proteins were identified by label-free mass spectrometry and data, analyzed by Progenesis QI software. (B) and (C) Functional clusters associated with bone marrow plasma proteins of patients diagnosed with myelodysplastic syndrome with ring sideroblasts (MDS-RS) and with myelodysplastic syndrome with excess blasts (MDS-EB), according to DAVID Functional Annotation Bioinformatics Analysis (https://david.ncifcrf.gov). Uniprot accession numbers were uploaded to DAVID platform and clusters were defined by enrichment scores and p-values [-log (p-value)]. Data refer to patients listed in . (D) and (E) Quantitative data, statistical analysis and gene symbols associated with in silico analysis of CEP55, TLN1 and MSN gene expression in CD34+ cell markers from MDS patients presented in the Gene Expression Omnibus (GEO) data bank (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE58831). This study evaluated gene expression from 159 patients with MDS patients and 17 healthy controls from bone marrow samples and microarray gene expression was performed using the Affymetrix GeneChip Human Genome U133 Plus 2.0 arrays.