Regulative Loop between β-catenin and Protein Tyrosine Receptor Type γ in Chronic Myeloid Leukemia.

Protein tyrosine phosphatase receptor type γ (PTPRG) is a tumor suppressor gene, down-regulated in Chronic Myeloid Leukemia (CML) cells by the hypermethylation of its promoter region. β-catenin (CTNNB1) is a critical regulator of Leukemic Stem Cells (LSC) maintenance and CML proliferation. This study aims to demonstrate the antagonistic regulation between β-catenin and PTPRG in CML cells. The specific inhibition of PTPRG increases the activation state of BCR-ABL1 and modulates the expression of the BCR-ABL1- downstream gene β-Catenin. PTPRG was found to be capable of dephosphorylating β-catenin, eventually causing its cytosolic destabilization and degradation in cells expressing PTPRG. Furthermore, we demonstrated that the increased expression of β-catenin in PTPRG-negative CML cell lines correlates with DNA (cytosine-5)-methyl transferase 1 (DNMT1) over-expression, which is responsible for PTPRG promoter hypermethylation, while its inhibition or down-regulation correlates with PTPRG re-expression. We finally confirmed the role of PTPRG in regulating BCR-ABL1 and β-catenin phosphorylation in primary human CML samples. We describe here, for the first time, the existence of a regulative loop occurring between PTPRG and β-catenin, whose reciprocal imbalance affects the proliferation kinetics of CML cells.


Introduction
Chronic myeloid leukemia (CML) is a myeloproliferative disease affecting approximately 1 per 200,000 persons per year in industrialized countries. Many treatment improvements have been achieved recently, especially in the development of new drugs, but a mortality rate of 2%-3% per year remains [1].
PTPRG. We confirmed that inhibition of BCR-ABL1 activity following treatment with imatinib mesylate in K562 cells causes the dephosphorylation of β-catenin Y654 and consequent protein degradation, as demonstrated by Gambacorti-Passerini's group in 2008 [18]. The interference with BCR-ABL1 activity by PTPRG is further supported by the finding that, in K562 γ1 cells, β-catenin and its Y654 phosphorylated form are strongly reduced in comparison with the mock control clone, making the evaluation of relative phosphorylation level challenging to assess ( Figure 1E). According to the data described above, the knock-down of PTPRG through RNA interference in unfractionated LAMA-84 cells ( Figure 1H) resulted in an evident increase of phospho-and total-β-Catenin, as demonstrated by Western blotting and immunofluorescence experiments ( Figure 1I,J).  were stably transfected with an empty vector pCR 3.1 (K562 mock) or with pCR 3.1 Protein tyrosine phosphatase receptor type γ (PTPRG) (K562 γ1) vector, containing the sequence of full-length PTPRG (PTPRG FL). qRT-PCR and Western Blotting show the fold change expression of PTPRG relative to K562 mock in transfected K562 cells. (B) Two clones from LAMA-84 cells, expressing alternatively low and high PTPRG levels, were selected and screened by qRT-PCR. Additionally, immunoprecipitation, combined with Western blotting analysis, confirmed PTPRG protein expression in these clones.
(C) Inhibition of PTPRG correlates with BCR-ABL1 tyrosine phosphorylation enhancement, and restored expression of total and phospho-β-catenin in K562 cells transfected with PTPRG FL. As expected, there was no response to the PTPRG inhibitor (two hours-treatment) in K562 transfected with the empty vector. (D) LAMA-84 clones expressing low or high PTPRG protein were treated with PTPRG IN for two hours. As anticipated, loss of PTPRG activity is associated with restored tyrosine phosphorylation and total β-catenin expression, only in the high expressing clone. (E) PTPRG-positive and negative K562 cells were treated with a specific inhibitor of BCR-ABL1, imatinib mesylate (IM), at 5 µM concentration for two hours. We observed, especially in combination with PTPRG transfection, marked dephosphorylation of β-catenin in parallel with its degradation. (F-G) Histograms representing densitometric quantification of Western blot signals from Figure 1C,D performed using ImageJ (U. S. National Institutes of Health, Bethesda, Maryland, USA, https://imagej.nih.gov/ij/, 1997-2018).
Only the two high-PTPRG conditions are statistically significant. (H-I) PTPRG down-regulation by RNA interference, through a specific siRNA (siPTPRG), as assessed by qRT-PCR and Western blot, and expressed as fold of increase. PTPRG down-regulation by siRNA is associated with the up-regulation of total and pY654β-catenin. The lack of signal in the line corresponding to total β-catenin is due to the different affinity of the individual antibodies (including secondary antibodies). The result is further confirmed by immunofluorescence analysis shown in figure (J) both pY654 β-catenin and total β-catenin protein (red) are increased in PTPRG (green) silenced cells. The scale bar length represents 10 or 20 µm. Down-regulation of PTPRG through transfection of specific siRNA in unfractionated LAMA-84 restores the detectable quantity of phospho-and total β-catenin. Pictures are representative of at least three experiments. Fold of increase in the graphics is the mean values of 3 biological replicates. p-value < 0.05 was considered statistically significant. Annotations for * p-value < 0.05, ** p-value < 0.01, *** p-value < 0.001, and **** p-value < 0.0001 are provided accordingly. Error bars indicate the SD for the three replicates.
In summary, in PTPRG-positive leukemic cells, the activity of this phosphatase causes the decrease of BCR-ABL1 and β-catenin tyrosine phosphorylation and the resulting degradation of the latter.

β-catenin and PTPRG Belong to the Same Protein Complex
Previous studies reported that BCR-ABL1 increases β-catenin stability through its phosphorylation activity in CML cells [18]. It can thus reasonably be expected that BCR-ABL1, β-catenin, and PTPRG belong to the same protein complex. So, we first evaluated the presence of β-catenin and PTPRG in a multi-molecular complex, by performing pull-down assay experiments. As a "bait" protein, we used the inactivated PTPRG substrate-trapping mutant (D1028A) [12] recombinant protein immobilized on nickel-agarose beads and challenged with protein extracts from K562 (BCR-ABL1 positive) and U937 (BCR-ABL1 negative) cell lines. This approach detects trapping of the substrate into the PTP catalytic pocket [20] and has allowed us to evaluate whether an interaction between PTPRG and β-catenin occurs. Co-precipitation happened in both cell lines, suggesting that this interaction does not require the expression of the oncogene BCR-ABL1 (Figure 2A). Co-immunoprecipitation experiments in cells in vivo expressing the substrate trapping form of PTPRG (K562 D1028A) confirm that β-catenin also binds PTPRG in a fully native condition ( Figure 2B).

PTPRG Expression Increases β-catenin Affinity Binding to Its Degradation Complex
The loss of β-catenin Y654 phosphorylation by BCR-ABL1 leads to the binding between β-catenin, and its multi-protein "destruction complex", that includes the tumor suppressors Axin1 [21] Axin1 carries binding sites for CK1 and GSK-3β and coordinates the whole phosphorylation events, involving β-catenin proteolysis by the 26S degradation machinery and acts as a tumor suppressor in hepatocellular carcinoma [21,22]. Immunoprecipitation experiments determined that, in K562 cells, β-catenin is present in a complex with Axin1 only in the presence of PTPRG ( Figure 2C). Besides, the restored expression of PTPRG in the K562 cell line (K562 γ1) induces a strong up-regulation of Axin1, along with the down-regulation of β-catenin, compared to PTPRG-negative K562 (K562 mock) ( Figure 2D).
As the additional confirmation of β-catenin-proteolysis events driven by PTPRG, we inhibited β-catenin degradation using the proteasome 26S subunit inhibitor, MG-132, in K562 cells expressing, or not, PTPRG. The proteasome inhibitor blocked the degradation of the protein, but dephosphorylation still occurred in the presence of active PTPRG ( Figure 2E), indeed suggesting that PTPRG dephosphorylates β-catenin. To further strengthen this finding, we treated the unfractionated PTPRG-positive LAMA-84 for two hours with the proteasome inhibitor MG-132 (10 µM) alone, and with the two inhibitors PTPRG IN (10 µM) and MG-132 (10 µM) combined, respectively. In keeping with the proposed mechanism, restored expression of the total β-catenin protein was associated with an increased level of Y654 phosphorylation in the sample treated with both inhibitors. Conversely, the blocking of the 26S proteasome without the inhibition of the phosphatase PTPRG prevents the degradation of the total protein but does not increase its phosphorylation level ( Figure 2F). In other words, the lowering of Y654 phosphorylation levels is not the result of β-catenin degradation, but of PTPRG activity driving β-catenin degradation.

PTPRG Down-Regulates β-catenin and Affects the Expression of Its Transcriptional Targets
We next evaluated whether β-catenin transcriptional targets were vicariously affected by PTPRG expression in leukemic cells, as should be expected by the previous findings.
TCF4/β-catenin complex is known to inhibit the transcription of the p21/WAF1 gene [23]. In keeping with this observation, we noticed that the higher expression of PTPRG is closely related to the reduced expression of β-catenin and consequent overexpression of p21/WAF1 in K562 cells ( Figure 3A), while the contrary was observed when PTPRG is down-regulated by a specific siRNA (siPTPRG) in LAMA-84 cells ( Figure 3B). Similarly, the up-regulation of PTPRG is also associated with the decreased transcription of two genes positively regulated by β-catenin and involved in cellular proliferation: MYC [24], and β-catenin itself, which acts as a co-transcription factor (with TCF4/LEF) on its promoter [25] ( Figure 3A). Opposite results were obtained when PTPRG is down-regulated in the PTPRG positive CML cell line (LAMA-84), further confirming the finding ( Figure 3B).
Differently from what was expected, the regulation of the Cyclin D1 gene, positively regulated by β-catenin, in our experimental conditions was not inversely related to PTPRG expression. The boost of Cyclin D1 expression is likely to be offset by the up-regulation of p21/WAF1, both at mRNA and protein levels ( Figure 3A-C), thus avoiding an unchecked activation of the cell cycle. Moreover, we demonstrated by Western blotting that, despite the fact that PTPRG increases the expression of Cyclin D1 mRNA, the protein is mostly localized in the cytosolic compartment in PTPRG positive K562 cells ( Figure 3C), thus preventing its activity as a cell cycle inducer in the nucleus.
Given these premises, we anticipate that PTPRG inactivation by PTPRG IN should have a positive impact on CML cell proliferation. To test this hypothesis, we performed dose-response experiments and evaluated the extent of BCR-ABL1 protein expression and Y245 phosphorylation. As expected, there is a dose-dependent increase of colony size and volume ( Figure 3D,E), that correlates with the enhancement of Y245 phosphorylation ( Figure 3F), thus confirming and extending our previous observation on PTPRG involvement in the regulation of CML cell proliferation [12]. Interestingly, the samples that showed the highest number and size of colonies (treatment with 0.1 and 0.2 µM PTPRG IN) were the same ones that expressed the highest level of phospho-BCR-ABL1, as demonstrated by Western blotting ( Figure 3F).

β-catenin Transcriptional Activity Correlates with DNMT1 Expression
Since we have previously demonstrated that PTPRG expression is linked to the promoter methylation levels [12], we hypothesize that DNMT proteins might negatively affect the transcription of this phosphatase. We focused on DNA (cytosine-5)-methyltransferase 1 (DNMT1), a DNA-binding enzyme responsible for the down-regulation of many tumor suppressor genes through hypermethylation of their promoter regions and a downstream effector of APC/β-catenin/TCF4 signaling [26]. It has been reported that the inhibition of β-catenin/TCF4 transcriptional activity, through the N-terminal deletion dominant-negative mutant, decreases DNMT1 mRNA levels [17]. Interestingly, DNMT1 transcript inversely correlates with PTPRG expression, as shown by the differential expression level in K562, LAMA84, and PTPRG-silenced LAMA84 ( Figure 4A,B). Moreover, in K562 cells expressing β-catenin (K562 mock), the β-catenin signal disruptor PNU-74654, which prevents β-catenin binding to TCF4/LEF transcriptional co-factor, leads to the down-regulation of DNMT1 mRNA ( Figure 4C).
Three active DNA methyltransferases have been identified in mammals: DNMT1, DNMT3a, and DNMT3b. DNMT1 is considered essential for the maintenance of DNA methylation, while DNMT3a and DNMT3b are involved in the first step of methylation [26]. We detected, by Western blotting, the higher expression of DNMT1 and DNMT3b in PTPRG-negative K562 cells compared to PTPRG expressing LAMA-84 cells, while DNMT3a, considered as a tumor suppressor in hematological malignancies [27], followed an opposite trend of expression, being present only in LAMA-84 cells ( Figure 4D).  silencing through a specific siRNA against DNMT1 as well as 5-azacytidine treatment (72 h) leads to a significative up-regulation of PTPRG mRNA in K562 cells. The combination of these two treatments induces a 5-fold increase of PTPRG mRNA expression, that becomes detectable at the protein level. Pictures are representative of at least three experiments. Fold of increase in the graphics is the mean values of 3 biological replicates. p-value < 0.05 was considered statistically significant. Annotations for * p-value < 0.05, ** p-value < 0.01, and *** p-value < 0.001 are provided accordingly. Error bars indicate the SD for the three replicates.

DNMT1 Binds PTPRG Promoter Region and Regulates PTPRG Expression
We therefore evaluated whether DNMT1 could represent a modulator of PTPRG expression through the regulation of its promoter methylation status. As demonstrated by previous works [28,29], DNMT1 can bind gene promoter regions to control their transcription by methylation. Furthermore, the PTPRG promoter region contains two distinct CpG islands available for methylation by DNMT1 and, consequently, potential binding sites for this protein.
Chromatin immunoprecipitation assay is a powerful technique suitable to map the interaction of proteins with specific genomic regions. This assay confirmed the interaction between DNMT1 and the CpG islands within the PTPRG promoter region ( Figure 4E), indicative of a role for DNMT1 in PTPRG down-regulation.
As further confirmation of our data, we sequenced the genomic region of the PTPRG promoter, after bisulfite conversion of the DNA from K562 and LAMA-84 cells. The sequencing analysis revealed a solid methylation pattern of this region only in cells with low expression of PTPRG ( Figure 4F,G), stronger in K562 cells compared with both clones (high and low-PTPRG expression) of LAMA-84.
In order to validate the involvement of DNMT1 in PTPRG transcriptional regulation, we treated PTPRG-negative K562 cells with a specific siRNA against DNMT1, alone or in combination with 5-azacytidine, an inhibitor of DNA methyltransferases ( Figure 4H). Although even a single treatment leads to a restored expression of PTPRG, the combination of 5-azacytidine with DNMT1 siRNA caused a 5-fold increment of PTPRG mRNA in comparison with control. Increased mRNA expression matches with the recovery of detectable protein levels ( Figure 4I,J).

PTPRG Dephosphorylates BCR-ABL1 and β-catenin in CML Primary Cells
We finally tested whether PTPRG affects BCR-ABL1 and β-catenin phosphorylation in primary CML cells. We first evaluated by qRT-PCR (data not shown), and flow cytometry the level of PTPRG in peripheral blood and bone marrow leukocytes obtained at the time of diagnosis from nine CML patients (Table 1) ( Figure 5A,B left panels). Five CML patients expressed detectable PTPRG levels, while four were PTPRG-negative. We tested the phosphorylation level of BCR-ABL1 in all nine patients and that related to phospho β-catenin in 6 out of 9 patients. In all the patients expressing PTPRG, 10µM PTPRG IN enhanced Y245 phosphorylation of BCR-ABL1 and, consistently with our previous results, Y654 phosphorylation of β-catenin ( Figure 5A,B right panel). Interestingly, two patients samples presenting undetectable levels of PTPRG (MFI 1.15, patient 3 and 1.03, patient 8) did not show increased BCR-ABL1 Y245 phosphorylation in response to PTPRG IN (Figure 5A-D). These results fully confirm and extend to primary cells the previous data that indicate an active role of this phosphatase in BCR-ABL1 activation and β-catenin degradation and, consequently, the inverse correlation between PTPRG expression and BCR-ABL1 activity in cells isolated by the peripheral blood of CML patients.

Discussion
Our previous works demonstrated the role of tumor suppressor for PTPRG in CML [12] and, in a different cell context, its capability to target JAK2 [30], a kinase governing pathways of relevant therapeutic interest in CML [31]. Here, we extend these previous findings by exploring in deeper detail the molecular mechanisms underlying the tumor-suppressive effect of this phosphatase. We proved that PTPRG dephosphorylates the residue Y245 in the BCR-ABL1 SH2-kinase-domain linker, essential for the full activation of this kinase and one of five high-confidence phosphorylation sites of BCR-ABL1, with structural and functional roles [32]. Dephosphorylation, with consequent BCR-ABL1 inactivation or weakening, represents a dramatic event for CML cells leading to increased apoptosis and differentiation, followed by reduced proliferation and aggressiveness [33]. This evidence is of considerable significance, thinking the role of Leukemic Stem Cells (LSC) in the resistance to Tyrosine Kinase Inhibitors and aberrant activation of self-renewal [34]. In the last few years, many studies have suggested the active contribution of β-catenin on LSC maintenance [35], even if the relationship between β-catenin and BCR-ABL1 is still under investigation. Some authors described BCR-ABL1 as a β-catenin target [36], demonstrating that β-catenin-deficient murine CML cells showed lower BCR-ABL1 protein level and phosphorylation activity. MYC, a β-catenin-downstream gene, controls BCR-ABL1 transcription, suggesting that BCR-ABL1 is an indirect target of β-catenin [37]. Furthermore, lncRNA-BGL3 was highly induced in response to the disruption of BCR-ABL1 expression, or by inhibiting BCR-ABL1 kinase activity in K562 cells and leukemic cells derived from CML patients. Conversely, BCR-ABL1 represses lncRNA-BGL3 expression through MYC-dependent DNA methylation [38]. Moreover, BCR-ABL1 regulates β-catenin through the phosphorylation of Tyr 86 and 654 that affects the stabilization of cytosolic β-catenin and its binding with its degradation complex [18]. Altogether, these data show that β-catenin, as Wnt pathway effector, is required to maintain normal HSC function, while the loss of this protein hampers CML progression [36]. Of note, nuclear β-catenin is involved in intrinsic BCR-ABL1 kinase-independent TKI resistance in primary CML cells [39] and PTPRG co-precipitates with β-catenin in enterocytes derived from transgenic mice, featuring the loss of the wild-type Apc allele (ApcMin−/−). In tumors derived from these mice, reduced levels of tyrosine-phosphorylated β-catenin have been reported [40].
Collectively, our data demonstrate that β-catenin degradation is controlled by PTPRG expression: PTPRG can dephosphorylate BCR-ABL1, thus preventing β-catenin tyrosine phosphorylation and, at the same time, apparently directly dephosphorylates β-catenin, causing the almost complete proteolysis within two hours ( Figure 6, left panel). Based on these experiments, we cannot assess whether the dephosphorylation of β-catenin is a consequence of BCR-ABL1 inhibition mediated by PTPRG, or the effect of a direct dephosphorylation. However, we have shown that PTPRG and β-catenin belong to the same complex, as they co-immunoprecipitate. This phenomenon is independent of BCR-ABL1 expression, as it occurs also in the BCR-ABL1-negative U937 cell line, strongly suggesting a direct interaction between PTPRG and β-catenin.
The loss of β-catenin reduces its transcriptional activity, as reflected by the down-regulation of MYC and β-catenin mRNA, in combination with p21/WAF1 mRNA up-regulation, which may explain lower cellular proliferation rate in the presence of PTPRG.
DNA methylation is a significant epigenetic modification in mammals, and, under physiological condition, 80% of the whole genome, not associated with promoter sequences, is methylated. On the contrary, in cancer cells, CpG-rich promoter methylation catalyzed by DNA methyl-transferases is a common mechanism of tumor suppressor silencing [41]. In our previous work, we demonstrated that the PTPRG promoter is methylated and that 55% of leukocytes from CML patients treated with TKI showed demethylation of the region and a re-expression of PTPRG [12].
DNMT1 is the major DNA-methyl-transferase responsible for the maintenance of DNA methylation after DNA replication; by contrast, DNMT3a and DNMT3b are de novo methyl-transferases that are often dysregulated in many types of tumors, such as colon, breast and pancreatic cancers, thus representing a therapeutic target [26]. DNMT3a has recently emerged as one of the most important tumor suppressors in hematological malignancies, in particular in AML cells, featuring a high number of mutations that are involved in resistance to some types of drugs [27]. DNMT1 and DNMT3b cooperate for tumor suppressor silencing [42], and we observed a higher expression of these two proteins in K562 cells, as compared to PTPRG-expressing LAMA-84. Besides, β-catenin/TCF4 transcriptional complex regulates DNMT1 expression. Indeed, it was demonstrated that the inactive dominant-negative version of this protein complex decreases DNMT1 expression [17].
We present here a cross-talk between DNMT1 expression and PTPRG promoter methylation, which was confirmed by the results of chromatin immunoprecipitation and by the evidence of increased PTPRG transcription in K562 following DNMT1 down-regulation. Figure 6. Schematic representation summarizing the regulative loop between β-catenin and PTPRG. In low-PTPRG expressing cells, BCR-ABL1 phosphorylates β-catenin in Tyrosine 86 and 654 in CML cells. This phosphorylation affects β-catenin cytosolic stabilization, blocking its binding with the destruction complex. This condition allows β-catenin to translocate into the nucleus and act as a transcriptional activator for many genes, including DNMT1 that is responsible for PTPRG promoter methylation and consequent down-regulation. On the other hand, in high-PTPRG expressing cells, PTPRG dephosphorylates BCR-ABL1, preventing β-catenin tyrosine phosphorylation; in addition, PTPRG directly dephosphorylates β-catenin, causing its proteolysis through the binding with its degradation complex. This condition affects the DNMT1 transcription, with the consequent down-regulation of this methyl-transferase, and hypo-methylation of PTPRG promoter region.
These results reveal a previously unrecognized regulative loop between β-catenin and PTPRG, with β-catenin controlling PTPRG transcription. At the same time, through DNMT1, β-catenin controls both BCR-ABL1 (through MYC expression) and PTPRG itself. On the other hand, PTPRG and BCR-ABL1 regulate β-catenin levels by modulating, in opposite ways, its tyrosine phosphorylation levels. Also, PTPRG increases the expression of Axin1, a key regulator of β-catenin degradation [15].
Of relevance is the observation that the same mechanism described in CML cell lines is active in leukocytes from CML patients, where PTPRG inhibition results in the up-regulation of Y245-phospho-BCR-ABL1 and Y654-phospho-β-catenin. Conversely, cells from CML patients expressing undetectable levels of PTPRG displayed the highest levels of Y245-phospho-BCR-ABL1 and Y654-phospho-β-catenin. Of note is the fact that they were refractory to PTPRG IN treatment, a key observation further confirming an exquisite specificity of this chemical inhibitor, utilized for the first time in cellular models.
Altogether this study describes a novel PTPRG-dependent regulative loop involving critical regulatory components of the BCR-ABL1 signaling pathway (Figure 6), further strengthening the role of PTPRG as a crucial tumor suppressor in CML cells and a possible candidate for specific therapies aimed at its re-expression or activation.

CML Patients
CML cells were collected from patients with untreated, chronic-phase disease, after informed consent (described in Table 1). 20 × 10 6 cells cryopreserved in 500 µL of 90% FBS and 10% DMSO were thawed using 4 µL of deoxyribonuclease I (260 U/µL), seeded in IMDM (Thermo Fisher Scientific, Monza, Italy) with 20% FBS for three hours and then treated with PTPRG IN, at different concentrations for two hours at 37 • C. Cells were lysed in boiling "sample buffer" (SB): 160 mM Tris, 20% glycerol, 5% β-mercaptoethanol, 4% SDS, 0.01% bromophenol blue and subjected to Western blotting. The study was approved by the Ethical Committee of Azienda Ospedaliera Universitaria Integrata Verona, protocol N. 1828, 12 May 2010, "Institution of cell and tissue collection for biomedical research in Onco-Hematology". The study was approved by the Local Ethics Committee, AOUI Verona (Permit Number: 25066); informed consent, in accordance with the declaration of Helsinki, was obtained from each patient.

Immunoprecipitation
Specific antibodies against β-catenin (3 µg, Santa Cruz Biotechnology Inc, Heidelberg Germany) and PTPRG [43] were added to 30 µL of Protein G-Sepharose beads (GE Healthcare Europe GmbH, Milan, Italy) and incubated for one hour at 4 • C with gentle rocking. Then, washed beads bound to the antibody were added to 500 µg of protein lysates and incubated again for three hours at 4 • C with gentle rocking. Finally, the beads were washed, resuspended in Sample Buffer 2×, denaturated at 95 • C and subject to Western blotting.

Pull-Down Assay
Intracellular domain (ICD) of PTPRG D1028A (PTPRG inactive mutant) and enhanced green fluorescent protein (eGFP) construct, both cloned in T7-based HisG-tagged vector expression pRSET A (Thermo Fisher Scientific, Milan, Italy), were expressed in BL21 (DE3) pLysS Escherichia coli (Thermo Fisher Scientific, Milan, Italy). For the pull-down assay, 3 mg of total bacterial protein lysates were conjugated to 30 µL of HIS-Select Nickel Affinity Gel (MilliporeSigma, Milan, Italy) and incubated, after extensive washing, with 500 µg of protein lysates from K562 and U937 cell lines for 3 h at 4 • C, with gentle rocking. The beads were collected, washed twice with lysis buffer and once with cold TBS, and then subjected to SDS-PAGE and Western blotting.

Methylation Analysis
The DNA extraction was performed using DNeasy Blood & Tissue Kit (Qiagen Sciences Inc, Germantown, MD, USA), according to the manufacturer's instruction. Subsequently, the DNA (2 µg) was subjected to bisulfite conversion and purification using EpiTect™ Fast DNA Bisulfite Kit (Qiagen Sciences Inc, Germantown, MD, USA), according to the manufacturer's protocol. After bisulfite conversion, the PTPRG promoter region was analyzed by PCR amplification and sequencing. PCR reaction was made using 50× Advantage ® 2 Polymerase Mix (Takara Bio USA Inc, Mountain View, CA, USA). The PCR product was purified with ExoSAP-IT™ PCR Product Cleanup Reagent (Thermo Fisher Scientific, Waltham, MA, USA), and subjected to Sanger sequencing. The analysis of sequencing was performed with Sequencer 5.4.6 software (Gene Codes Corporation, Ann Arbor, MI, USA) to evaluate the C to T conversion rate.

Statistical Analysis
The data analysis was performed using GraphPad ® 8.3.0 Instat software (GraphPad Software, San Diego, CA, USA). The Student's two-tailed unpaired t-test was applied to qRT-PCR and Colony assay (number and volume of colonies) experiments. For qRT-PCR analysis, we considered absolute values, even though the graphs show just the fold-change, representing the data more clearly. Also, the student's two-tailed unpaired t-test was applied to the densitometric comparisons depicted in Figure 1F,G, Figure 5C,D. Each PTPRG IN condition was compared to the DMSO control. The analysis was done separately for each different protein employing ImageJ software (U. S. National Institutes of Health, Bethesda, MD, USA https://imagej.nih.gov/ij/, 1997-2018). Results with a p-value < 0.05 were considered statistically significant.