Epigenetic DNA changes in childhood acute lymphoblastic leukemia - the drug treatment effect.

The active DNA demethylation process may be linked with aberrant methylation and may be involved in leukemogenesis. We investigated the role of epigenetic DNA modications in childhood acute lymphoblastic leukemia (ALL) development. We analyzed the levels of 5-methyl-2′-deoxycytidine (5-mdC) oxidation products in the cellular DNA and urine of children with ALL (at diagnosis and after completion of chemotherapy, n=55) using two-dimensional ultra-performance liquid chromatography with tandem mass spectrometry (2D UPLC–MS/MS). Moreover, the expression of Ten Eleven Translocation enzymes (TETs) at the mRNA and protein levels was determined. Additionally, the ascorbate level in blood plasma was analyzed. Before treatment, the ALL patients had profoundly higher levels of the analyzed modied DNA in their urine than the controls. After chemotherapy, we observed a statistically signicant decrease in active demethylation products in urine, with the nal level similar to the level characteristic of healthy children. The level of 5-hmdC on the DNA of the blood leukocytes in the patient group was signicantly lower than it was in the control group. Our data suggest that urinary excretion of epigenetic DNA modication may be a marker of pediatric ALL pathogenesis and a reliable marker of chemotherapy response. 8-oxo-7,8-dihydro-2′-deoxyguanosine; 5-mdC: 5-methyl-2′-deoxycytidine; 5-hmdC: 5-(hydroxymethyl)-2′-deoxycytidine; 5-fdC: 5-formyl-2′-deoxycytidine; 5-cadC: 5-carboxy-2′-deoxycytidine; 5-hmdU: 5-(hydroxymethyl)-2′-deoxyuridine; MLL: mixed lineage leukemia; AML: acute myeloid leukemia; LC/GC–MS: liquid chromatography followed by gas chromatography with isotope dilution mass spectrometric detection; UPLC ‐ UV: ultra-performance liquid chromatography coupled with UV detection; BER: base excision repair; MMR: DNA mismatch repair; NER: nucleotide excision repair; RT-qPCR: quantitative real-time PCR


Introduction
Acute lymphoblastic leukemia (ALL) is the most common pediatric malignancy. ALL can develop from any lymphoid cell blocked at a particular stage of development and usually originates in a single B-or Tlymphocyte progenitor 1 . Most patients diagnosed with ALL have genetic aberrations that contribute to a higher rate of proliferation and impaired differentiation of lymphoid hematopoietic progenitors 2 .
One of the reasons for the observed genetic alterations may be aberrant DNA methylation that leads to changes in the expression of hematopoietic genes. Moreover, in several studies, it has been observed that the cytosine methylation pattern of leukemic cells differ from that of nonleukemic cells [3][4][5] .
Both 5-fCyt and 5-caCyt are subsequently removed by base excision repair (BER) or nucleotide excision repair (NER) and replaced by cytosine to complete the process of active DNA demethylation 8 . 5-Hydroxymethyluracil (5-hmUra) may also be generated by TET enzymes and removed via the BER pathway. The eliminated bases (or deoxynucleosides) are released into the bloodstream and eventually are also found in urine 9 .
It is likely that active DNA demethylation is linked with aberrant methylation and may be involved in leukemogenesis. Indeed, hematological malignancies were among the rst in which aberrant demethylation status was discovered, and the TET gene was initially de ned as a fusion partner of the mixed lineage leukemia (MLL) gene in acute myeloid leukemia (AML). Importantly, TET inactivation leads to the dramatic deregulation of hematopoiesis that, in turn, triggers blood malignancies 10 .
Quanti cation of epigenetic DNA modi cations may elucidate the consequences of TET expression de ciency. The functional consequences of epigenetic changes may be critical for disease initiation and progression. However, changes in the level of epigenetic DNA modi cations may also mirror environmental changes linked with drug delivery.
We thus sought to determine an array of endogenously generated DNA base modi cations (5-hmUra, 5-mCyt, 5-hmCyt, 5-fCyt, 5-caCyt), with the goal of determining whether these modi cations can serve as predictive biomarkers of leukemia disease process. TET enzyme expression at the mRNA and protein levels was also studied. Since ascorbate may substantially enhance the generation of 5-hmCyt (as well as other DNA epigenetic modi cations), probably by acting as a cofactor of TETs during the hydroxylation of 5-mCyt, its level was analyzed in the blood of the experimental groups.
Oxidative stress and impaired repair of oxidatively modi ed DNA have been documented for children with ALL 11,12 . Therefore, we also analyzed the 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodG) level (an established biomarker of oxidative stress/DNA damage) in the cellular DNA and urine of the ALL patients and the control group participants.

Results
Levels of epigenetic modi cations and 8-oxoGua in urine and DNA The ALL patients before treatment had substantially higher levels of analyzed modi cations in their urine than the controls, and the differences were signi cant for all compounds (Table 1). After chemotherapy, the patients exhibited a statistically signi cant decrease in active demethylation products in their urine (5-hmCyt, 5-hmdC, 5-mdCyt, 5-caCyt, 5-hmUra) both 33 days after starting the treatment (point B) and six months after the start of therapy (point C). Notably, the levels of analyzed compounds at point C were similar to those observed in the control group ( Figure 2). Signi cantly elevated levels of 8-oxodG were noted in the urine of patients after chemotherapy (2.8 nmol/mmol creatinine at point A and 5.7 and 4.2 at points B and C). The level of 5-hmdC in the DNA of blood leukocytes in the patient group was signi cantly lower than that in the control group (median value: 0.065 vs. 0.078, p=0.007). An inverse relationship was observed in the case of 8-oxodG, for which the ALL patients showed signi cantly higher levels than the controls (1.59 vs. 1.38, respectively, p=0.009, Table 2). Description of initial diagnosis, based on immunophenotypes and genetic aberrations was placed in Supplementary Materials (Table  S1). The ALL patients were divided based on the subtype of disease (B and T type) and all the analyzed parameters in urine were determined separately for this subtypes (Table S2 in Supplementary Materials).

Expression of TET and TDG mRNA
The expression of TET1 in the ALL patients was signi cantly higher than that in the control. The expression of TET2 and TET3 in the individuals with ALL was signi cantly lower than that in the controls (p = 0.005 and p= 0.00004, respectively). A similar statistically signi cant difference was found for the expression of TDG in the ALL group and the control group (p=0.00001, Table 3).  Table 3. Expression of TET1, TET2, TET3 and TDG mRNA in the healthy controls and ALL patients ( * p<0.05).

Protein expression analysis
The protein expression of TET1 and TDG in the ALL group was signi cantly lower than that in the control group (p=0.007). In the case of the TET2 and TET3 proteins, we found no statistically signi cant differences ( Table 4). The protein expression of TDG signi cantly decrease after treatment (median value 1.808 vs 1.413 respectively, p = 0.006, Figure S1 in Supplementary Materials). Interestingly, signi cant correlations between TETs ("producer") and TDG ("eraser") proteins were found ( Figure S2-S5 in supplementary materials). A moderate positive correlations were observed between active DNA demethylation products and TET proteins in selected cell populations ( Figure S6 in supplementary materials).

Differences in the plasma concentrations of ascorbate
Plasma concentrations of ascorbate in the ALL patients were lower than they were in the controls (median value 62.4 vs 97.8 µM respectively, p = 0.00027, Figure 3).

Discussion
Acquiring an accurate measurement of 5-fdC, 5-cadC and 5-hmdU is challenging because not only are these modi ed bases in mammalian genomes expressed at very low levels but also the level of 5-hmdC is 3-4 orders of magnitude higher than the levels of the other forms of modi ed DNA, which affects the detection and quantitation of the less common metabolites. To address this di culty, we recently developed a rapid, highly sensitive and speci c isotope-dilution method of automated two-dimensional ultra-performance liquid chromatography with tandem mass spectrometry (2D-UPLC-MS/MS) that is speci cally tailored to analyze global levels of 5-mC, 5-hmC, 5-fC, 5-caC and 5-hmU simultaneously in cellular DNA and urine 16,17 . Importantly, many previous studies have used less reliable semiquantitative immunohistochemical methods to assess the levels of 5-hmC, 5-fC and 5-caC, and these results of these previous studies have lower reliability because of the low levels of modi cation in genomic DNA 18 .
Using the abovementioned methodology, we found a signi cant decrease in 5-hmdC in the leukocyte DNA of the ALL patients compared with the level in the healthy subjects. A similar decrease in this epigenetic modi cation has been found in many types of human malignancies [18][19][20][21] . One possible reason for the loss of 5-hmCyt is the decreased expression of TETs 19,22 . In agreement with this supposition, we found profound decreases in TET3 and TET2 mRNA expression in the leukocytes in the ALL patients. However, at the protein level, the only signi cant difference in the expression between the examined tissues was found for the case of TET1, which was signi cantly more abundant in the control group than it was in the patients. This difference may be linked with posttranslational context-dependent acetylation and deacetylation of TET lysine residues, which may enhance or reduce TET stability and activity 23,24 . Likely reason of the changes of TETs mRNA expression is methylation of CpG sites of promoter region which was found to correlate with gene expression changes in primary ALL cells 13 .
Another factor that in uences the activity of TET proteins is vitamin C. Several previous studies demonstrated that ascorbate may enhance the generation of 5-hmCyt in cultured cells and under in vivo conditions [25][26][27] . Ascorbic acid (AA) enhances the activity of TETs, likely at the active site by reducing Fe 3+ to Fe 2+ , which enables the donation of the electron needed to split an O-O into hydroxyl groups. Interestingly, the ALL patients analyzed in our study had signi cantly lower serum AA concentrations ( Figure 3). Thus, a shortage of vitamin C may be another factor critical to the observed decrease in the 5-hmC level in the leukocyte DNA of the patient group (in addition to decreased TET expression).
In ALL group, patients with ETV6/RUNX1 gene fusion (also known as TEL/AML1) have higher levels of analyzed modi cations in urine and for 5-hmdC, 5-hmCyt, 5-fCyt and 5-hmUra the differences were signi cant (Table S3 in Supplementary Materials). Interestingly, a bias towards hypermethylation was found of the ETV6/RUNX1-positive pediatric ALL patients 4,28 . In the case of hyperdiploidy no differences between the subtypes have been found.
In patients treated with chemotherapy (methotrexate or epirubicin), the 8-oxodG level in leukocyte DNA and in urine was increased almost two fold compared to that of the same patient group before treatment. One of the components of the therapy drug cocktail is the anthracycline derivative epirubicin. The cytotoxicity of these drugs has been attributed to the inhibition of topoisomerase II and the intracellular production of free radicals. In our previous study, as in this study, signi cant increases in the amount of oxidatively modi ed DNA bases, including 8-oxoGua, compared to the control levels, was found in the lymphocytes of cancer patients treated with epirubicin 29 .
Urinary excretion rates of the majority of the DNA epigenetic modi cations were signi cantly higher in the patient group than in the healthy subject group (Table 1). Recently, we and others found signi cant changes in urinary DNA epigenetic modi cations only in 5-hmdC from patients with solid tumors and lymphoma patients 17,30 . This result may indicate the essential effects of epigenetic changes of DNA on the malignant transformation observed in pediatric ALL (see also below).
Importantly, a profound decrease in these modi cations, to the level observed in the urine of the healthy subjects, was found after the completion of chemotherapy ( Figure 2). The likely reason for this decrease is the massive death of cells with a high proliferative rate, namely, cancer cells. In addition, cancer cells have a higher turnover of epigenetic marks. Therefore, the abovementioned decrease may be linked with the recovery of the patients. Moreover, one of the components of the ALL treatment, methotrexate (MTX), acts by inhibiting dihydrofolate reductase, which is critical for the reduction in methyltetrahydrofolate and for methylation during DNA replication, which in turn directly reduces cell proliferation 31 . Furthermore, MTX is directly involved in the inhibition of DNA methylation in pediatric leukemia patients 32 .
Normal cell development requires the tight control of the DNA methylation pattern, which is involved in the regulation of gene expression. The methylation of cytosine, usually in CpG dinucleotides, is a key epigenetic modi cation that exerts a profound impact on gene repression, cellular identity and the maintenance of genome stability 33 . As mentioned in the Background, equally important for DNA methylation arrangement is the opposite reaction, namely, active DNA demethylation, which can result in activation of previously silenced genes. This process is directly linked with DNA repair and the removal of epigenetic modi cations such as 5-fCyt and 5-caCyt (also 5-hmUra), which are replaced by unmodi ed cytosine. The removed bases (or deoxynucleosides) are released into the bloodstream and eventually appear in the urine 34 . Therefore, the most plausible source of the modi cations excreted in the analyzed urine is the DNA repair process. Both 5-FCyt and 5-caCyt may inhibit DNA replication, which may result in genome instability 35,36 . Therefore, effective enzymatic systems are involved in the removal of modi cations from DNA. Thus, TDG has robust excision activity toward 5-fCyt and 5-caCyt [37][38][39] . The activity of the abovementioned enzymes may contribute to the presence of modi ed bases in urine. Importantly, the expression of TDG mRNA was sharply increased in the ALL patients compared with controls (Table 3).
Another source of epigenetic marks excreted in urine, as analyzed in our study, may be the processive demethylation pathway, as proposed by Franchini et al. 40,41 . That is, 5-fCyt, 5-caCyt and 5-hmUra initiate processive DNA demethylation, which is primarily a single initiating event (such as a certain mismatch) that may trigger the processive demethylation of numerous 5-mCyts (and perhaps 5-hmCyts) on the same locus via the long-patch BER, DNA mismatch repair (MMR) or nucleotide excision repair (NER) pathways. Recent studies have demonstrated that 5-hmUra may trigger the removal of distant epigenetic modi cations (5-mCyt and 5-hmCyt) through the MMR-and long-patch BER-dependent pathways 42 . These ndings may explain the presence of 5-hmCyt and 5-mCyt deoxynucleosides in urine. Lesioncontaining oligomers in the NER/MMR/long-patch BER pathways may be the subject of intra/extracellular 5'-3' exonucleolytic digestion, which eventually results in the synthesis of 6-to 7nucleotide-long oligomers. However, the latter may be further degraded, and this poorly characterized post-excision processing may be eventually re ected in a modi ed deoxynucleotide yield 43 .
Interestingly, the urinary excretion rates found in this study for the control group/healthy children were similar to those we obtained for healthy adults, as described in our recently published study 17 . These ndings seem to re ect basic processes that govern metabolic pathways critical for maintaining certain patterns of epigenetic DNA marks.
Several recently published studies have demonstrated that certain epigenetic alterations are essential for leukemic transformation 13 . The results of the aforementioned studies demonstrated the hypermethylation of GC islands in all subtypes of ALL. However, in these ALL cells, genome-wide hypomethylation was also frequently detected. Moreover, the ALL genomes were characterized by few somatic mutations, while the depletion of epigenetic modi cations was simultaneously exacerbated in the majority of the pediatric ALL patients 13 . It is likely that profound changes in urinary excretion rates of epigenetically modi ed DNA in the ALL patients, in comparison with healthy children, as observed in this work, directly re ect the abovementioned dynamic changes in methylation signatures during disease development.
As explained above, a large body of evidence suggests that the level of 5-hmCyt in many human malignancies is substantially reduced [18][19][20][21] . Moreover, recently published data suggest that the decrease in the 5-hmCyt level may serve as a biomarker of early malignant transformation and can be used as a prognostic factor for patients with malignancies 21,44 . Consequently, the urinary modi cations described herein may be used as potential risk and response markers. Owing to di culties in obtaining specimens/tissues, such as bone morrow, especially from pediatric patients, the determination of epigenetic DNA modi cations in human urine may serve as an attractive noninvasive diagnostic option. Furthermore, the noninvasiveness of the test constitutes a strong argument for its application to largescale basic research and clinical studies dealing with the role of active demethylation in carcinogenesis.
We found a highly signi cant increase in epigenetic DNA modi cations in the urine of pediatric ALL patients compared with the level in healthy subjects. This increase likely re ects dynamic epigenetic changes in ALL pathogenesis. Importantly, after chemotherapy, urinary excretion rates of these modi cations returned to the level characteristic of healthy children, a nding that may be linked with patient recovery. Our results suggest that the urinary excretion of epigenetically modi ed DNA may be a marker of pediatric ALL pathogenesis and a reliable marker of chemotherapy response. Since ETV6/RUNX1 fusion is favorable prognostic marker our results, which demonstrated direct link between the analyzed modi cations in urine and ETV6/RUNX1 status, are a kind of support for the aforementioned statement.  The determination of the epigenetic modi cations and 8-oxodG levels in urine.

Materials And Methods
Two-dimensional ultra-performance liquid chromatography with tandem mass spectrometry (2D UPLC-MS/MS) was used for the epigenetic modi cation analysis of urine samples (with the exception of 5-hmUra). Urine samples were spiked with a mixture of internal standards at a 4:1 volumetric ratio. The 2D-UPLC−MS/MS system consists of a gradient pump and autosampler for one-dimensional chromatography, and a gradient pump and tandem quadrupole mass spectrometer with a UNISPRAY ion source was used for two-dimensional chromatography. Both systems were coupled with a column manager equipped with two programmable column heaters and two 2-position 6-port switching valves.
The at-column dilution technique was used between the rst and the second dimensions to improve the retention in the trap/transfer column. The sample molecules were then adsorbed to the packing material as very narrow bands that could be eluted with well-resolved, small-volume peaks. A diluting stream of water (0.5 mL/min) was pumped with a Waters 515 isocratic pump and mixed with the rst-dimension column e uent using a UPLC low-dead-volume tee valve. The following columns were used: CORTECS UPLC T3 Column (1.6 µm, 3 mm × 150 mm) with a CORTECS T3 VanGuard precolumn (1.6 µm, 2.1 mm × 5 mm) for the rst dimension, a Waters ACQUITY UPLC CSH C18 (1.7 µm, 2.1 mm × 100 mm) for the second dimension, and a Waters XSelect CSH C18 column (3.5 µm, 3 mm × 20 mm) as the trap/transfer column. The chromatographic system was operated in heart-cutting mode, which means that selected portions of e uent from the rst dimension were loaded onto the trap/transfer column by 6-port valve switching, which served as an "injector" for the second dimension of the chromatography system. Mass spectrometric detection was conducted with a Waters Xevo TQ-S tandem quadrupole mass spectrometer equipped with a UniSpray ionization source. The following common detection parameters were used: source temperature, 150°C; nitrogen desolvation gas ow, 1000 L/h; nitrogen cone gas ow, 150 L/h; desolvation temperature, 500°C; and nebulizer gas pressure, 7 bar. Collision-induced dissociation was obtained with argon (6.0 at 3 × 10-6 bar pressure) as a collision gas. The instrument response to all compounds was optimized by the infusion of 10 µM genuine compounds dissolved in water (10 µL/minute) in the mobile phase A stream via the mass spectrometer uidics system operating in the "mixed" mode using MassLynx 4.1 software IntelliStart feature. The quantitative and qualitative transition patterns and the speci c settings of the detector are summarized in Table S5. The chromatographic system was operated with MassLynx 4.1 software from Waters. Quantitative analyses were performed using the TargetLynx application. All samples were analyzed with three to six technical replicates. Due to the low sensitivity of the method used, the level of 5-hmUra was determined by highperformance liquid chromatography for pre-puri cation followed by gas chromatography with isotope  separation was performed with a Waters ACQUITY 2D-UPLC system with a photodiode array detector for the rst dimension of the 2D-chromatography (used for quanti cation of the unmodi ed deoxynucleosides and 5-mdC) and a Xevo TQ-S tandem quadrupole mass spectrometer (used for the second dimension of the 2D-chromatog nraphy to analyze the compounds from the rst dimension in positive mode, namely, 5-hmdC and 8-oxodG, to assure better ionization at higher acetic acid concentrations). The at-column dilution technique was used between the rst and second dimensions to improve the retention of the trap/transfer column. The following columns used: a Waters CORTECS T3 column (150 mm×3 mm, 1.6 µm) with a precolumn for the rst dimension, a Waters XSelect C18 CSH (100 mm×2.1 mm, 1.7 µm) for the second dimension and a Waters XSelect C18 CSH (20 mm×3 mm, 3.5 µm) column as a trap/transfer column. The chromatographic system was operated in heart-cutting mode, indicating that selected fractions of the e uent from the rst dimension were loaded onto the trap/transfer column by 6-port valve switching, which served as the "injector" for the second dimension of the 2D-chromatography process. The ow rate for the rst dimension was 0.5 mL/minute, and the injection volume was 2 µL. Separation was performed with a gradient elution for 10 minutes using a mobile phase of 0.05% acetate (A) and acetonitrile (B) (0.7-5% B for 5 minutes, column washing with 30% acetonitrile and re-equilibration with 99% A for 3.6 minutes). The ow rate for the second dimension was 0.3 mL/minute. The separation was performed with a gradient elution for 10 minutes using a mobile phase of 0.01% acetate (A) and methanol (B) (1-50% B for 4 minutes, isocratic ow of 50% B for 1.5 minutes, and re-equilibration with 99% A until the next injection). All samples were analyzed with three to ve technical replicates, of which the technical mean was used for further calculation. Mass spectrometric detection was performed using a Waters Xevo TQ-S tandem quadrupole mass spectrometer equipped with an electrospray ionization source. Collision-induced dissociation was obtained using argon 6.0 at 3 × 10 -6 bar pressure as the collision gas. Transition patterns for all the analyzed compounds and the speci c detector settings were determined using the MassLynx 4.1 IntelliStart feature set in quantitative mode to ensure the best signal-to-noise ratio and a resolution of 1 at MS1 and 0.75 at MS2.

Determination of ascorbate level in blood plasma by UPLC-UV
The analyses were performed using a method described earlier by Starczak et al. with some modi cations 48 .
Gene expression analysis RNA isolation was performed using the PAXgene blood RNA kit (Qiagen) following standard procedures. The concentration and purity of the RNA aliquots were veri ed spectrophotometrically with a NanoDrop 2000 spectrophotometer (Thermo Scienti c). The A 260 /A 280 ratio was used as an indicator of protein contamination, and the A 260 /A 230 ratio was used as a measure of contamination with polysaccharides, phenol and/or chaotropic salts. The quality and integrity of the total RNA were assessed by visualization of the 28S/18S/5.8S rRNA band pattern in a 1.2% agarose gel. Nondenaturing electrophoresis was carried out at 95 V for 20 minutes in TBE buffer (Tris, boric acid and EDTA). The gel was stained with ethidium bromide or SimplySafe and visualized using a G:BOX EF gel documentation system (Syngene).
The samples with RNA concentrations greater than 50 ng/μL were quali ed for further analysis. A total of 0.5 micrograms of total RNA from each sample (in a 20-μl volume) was used for cDNA synthesis by reverse transcription with a high-capacity cDNA reverse transcription kit (Applied Biosystems, catalog no. 43-688-14) according to the manufacturer's instructions. The reaction was carried out with a Mastercycler nexus gradient thermocycler (Eppendorf). To exclude contamination with genomic DNA, reverse transcriptase reaction also included a negative control. cDNA was either immediately used for qPCR or stored at −20°C.
Gene transcripts were analyzed by relative quantitative RT-PCR (RT-qPCR) with relevant primers and short hydrolysis probes substituted with locked nucleic acids from the Universal Probe Library (UPL, Roche - Table 6). The probes were labeled with uorescein (FAM) at the 5'-end and with a dark quencher dye at  TDG 5'-GAATGGAAGCGGAGAACG-3' 5'-TTGCTGTTCATTCACAACTGC-3' #41 Table 6. Primers and short hydrolysis probes used for the target gene mRNA expression analysis.

Protein expression analysis
Because of the lack of commercially available anti-TETs (Tet1, Tet2, Tet3) antibodies recommended for direct ow cytometry method, it was decided to use an indirect staining consisting of 2 steps involving labeling with unconjugated primary and compatible uorochrome-conjugated secondary antibodies. Both type of antibodies were titrated.
This process enables the determination of the antibody amount and concentration, resulting in the highest possible signal from the positive population and the lowest signal from the negative population.
Nonspeci c antibody-binding was eliminated by optimizing the amount and concentration of the antibody 50 .
Blood specimens were collected by venipuncture into tubes containing tripotassium EDTA and the cellular surface antigen-stabilizing agent TransFix. Before the staining procedure was performed, the cells were counted using a LUNA-II automated cell counter. This number was important for ensuring the correct amount of antibodies was added. For the purposes of our experiment, a volume of blood containing 100 000 cells was added to a 5-ml round-bottom tube and suspended in PBS to obtain a nal volume of 100 µl.
The samples were incubated at room temperature in the dark for 30 minutes. Then, the erythrocytes were lysed with 2 ml of FACS lysing solution (BD Biosciences) and incubated for 15 minutes under the same conditions as the previous incubation. After two washes with PBS and centrifugation at 500 × g for 5 minutes, the cell pellets were resuspended in 500 µl of BD Cyto x/Cytoperm™ solution (BD Biosciences) to x and permeabilize the cells. The cells were incubated in the dark for 20 minutes at room temperature.
Then, the cells were centrifuged at 500 x g for 5 minutes again. Cell pellets were resuspended in freshly prepared Perm/Wash Buffer and incubated for 10 minutes in the dark. Then, the cells were centrifuged, the supernatant was removed, and the pellets were treated by primary antibodies to the scheme shown in The secondary antibody solution of donkey anti-rabbit Alexa Fluor 488 or rabbit anti-goat Alexa Fluor 647 (Abcam) was prepared and 100 µL of the 1000-fold diluted antibody was added to the samples. One hundred microliters of Perm wash buffer was added to tube no. 1. Sample no. 2, which included only surface and secondary antibodies, was a control for the investigated proteins. The tubes were incubated at room temperature for 30 minutes in the dark. Then, 1 mL of Perm wash buffer was added per tube and incubated for 10 minutes at room temperature in the dark. Next, the samples were centrifuged (500 × g for 5 minutes), and the supernatant was removed by pipette. The cell pellets were resuspended in 250 µL of PBS and then analyzed using a three-laser BD FACS Canto II ow cytometer (BD Biosciences) 51 . The instrument was set up using a BD Cytometer Setup and Tracking beads. Data on 30,000 cells were acquired with BD FACSDiva software and analyzed with FlowJo 7.5.5 software (Tree Star Inc., Ashland, Oregon, USA).
Cell doublets were rejected by setting the gate for single cells in the plot linear forward-scatter area (FSC-A) vs. linear forward-scatter height (FSC-H). The leukocytes (granulocytes, monocytes and lymphocytes, and blasts) were gated on the basis of their morphological features, as measured by linear side-scatter area (SSC-A) and CD45 antigen expression (logarithmic CD45 Brilliant Violet 421-H). For each population, the uorescence was detected separately and plotted in histogram mode: counts vs. logarithmic Alexa Fluor 488-H and counts vs. logarithmic Alexa Fluor 647-H. The geometric mean of the uorescence intensity (GMFI) was calculated for the total cell population in each channel. The expression of the investigated proteins was calculated as the geometric mean uorescence intensity in the test sample as a fold change over the negative control (tube no. 2) (ratio= GMFI test tube/GMFI control tube).

Statistical analysis
The results are presented as median values, interquartile ranges and non-outlier ranges. Statistical The datasets obtained in the current study are available from the corresponding author on reasonable request.