Predictive value of tyrosine phosphatase receptor gamma for the response to treatment tyrosine kinase inhibitors in chronic myeloid leukemia patients

Protein tyrosine phosphatase receptor gamma (PTPRG) is a member of the receptor-like family protein tyrosine phosphatases and acts as a tumor suppressor gene in different neoplasms. Recent studies reported the down-regulation of PTPRG expression levels in Chronic Myeloid Leukemia disease (CML). In addition, the BCR-ABL1 transcript level is currently a key predictive biomarker of CML response to treatment with Tyrosine Kinase Inhibitors (TKIs). The aim of this study was to employ flow cytometry to monitor the changes in the expression level of PTPRG in the white blood cells (WBCs) of CML patients at the time of diagnosis and following treatment with TKIs. WBCs from peripheral blood of 21 CML patients were extracted at diagnosis and during follow up along with seven healthy individuals. The PTPRG expression level was determined at protein and mRNA levels by both flow cytometry with monoclonal antibody (TPγ B9-2) and RT-qPCR, and BCR-ABL1 transcript by RT-qPCR, respectively. PTPRG expression was found to be lower in the neutrophils and monocytes of CML patients at time of diagnosis compared to healthy individuals. Treatment with TKIs nilotinib and Imatinib Mesylate restored the expression of PTPRG in the WBCs of CML patients to levels observed in healthy controls. Moreover, restoration levels were greatest in optimal responders and occurred earlier with nilotinib compared to imatinib. Our results support the measurement of PTPRG expression level in the WBCs of CML patients by flow cytometry as a monitoring tool for the response to treatment with TKIs in CML patients.


PE-CY7
A tandem fluorochrome consisting of R-phyco erythrin BV Chronic myeloid leukemia (CML) accounts for 15-20% of all hematological malignancies amongst adults 1 . BCR/ ABL is the first oncoprotein discovered in patients with CML and the first target for therapeutic intervention with small molecule tyrosine kinase inhibitors (TKIs) based drugs 2 . To date, five different BCR-ABL TKIs (i.e., Imatinib Mesylate (IM), Nilotinib, Dasatinib, Ponatinib, and Bosutinib) have been approved by the United States Food and Drug Administration (FDA) for the treatment of patients with CML 3 . However, primary or secondary resistance to treatment with TKIs occur in some patients with CML, highlighting the need for better understanding the mechanisms of resistance and identification of the predictive biomarkers for the response to treatment 4 . Protein Tyrosine Phosphatase Gamma (PTPRG) is a tumor suppressor and a member of a family of receptors tyrosine phosphatases 5 . It has the ability to remove a phosphate group from the phosphorylated amino acid tyrosine that is present on its substrate protein, thus balancing the ABL1 tyrosine kinase activities 6 . PTPRG is located on the short arm (3p14.2) of chromosome 3 and consists of 30 exons. It has an extracellular domain, a transmembrane domain, a carbonic anhydrase-homologous amino terminus followed by a fibronectin III domain, and two intracellular PTPase catalytic domains of which the only one is active 7 . PTPRG regulates hematopoietic differentiation processes and is expressed in human hematopoietic precursor cells, as well as in neutrophils and monocytes 8 .
In 2010, Della Peruta et al. reported down-regulation PTPRG at both mRNA and protein levels in leukocytes of CML patients 9 . The status of PTPRG has also been reported to be important in the response to treatment with TKIs in CML patients 10 . In another study, our group identified a single Nucleotide Polymorphism (SNPs) (rs62620047) in PTPRG (Y92H) in patients who failed Imatinib Mesylate (IM) treatment 11 . More recently, hypermethylation of PTPRG loci was reported to be a molecularly independent mechanism of resistance to treatment with TKI in CML patients 12 .
Flow cytometry is a simple and robust technique employed in many laboratories for the diagnostic and research purposes. It can also be of potential value in the monitoring of CML patients and their response to therapeutic intervention [13][14][15][16] . In a previous study, we described the production of a monoclonal antibody that recognizes the extracellular domain of PTPRG (TPγ B9-2) on CML cell lines and patients' samples. This is a unique antibody with a potential capacity for monitoring the PTPRG expression in CML patients 17 .

Results
To our knowledge, mAb TPγ B9-2 is currently the only mAb for use in the detection of PTPRG protein by flow cytometry. Therefore, in this study, the changes in the expression level of PTPRG protein were determined in the WBCs of seven healthy individuals and 21 CML patients at the time of diagnosis and following treatment with BCR-ABL TKIs, using flow cytometry.
Characteristics of CML patients at the time of diagnosis, their responses to the treatment with TKIs and healthy control participants. Out of the 21 CML patients examined in this study, 18 (86%) were diagnosed at chronic phase (CP) and 3 (14%) at accelerated phase (AP) 18 . The mean age of the 21 CML patients was 38.21 years and those of seven healthy controls 35.2 years. Eleven patients were treated with Imatinib (400 mg) and of these only two patients had optimal responses. Optimal response has also been developed in one patient following treatment with 600 mg of Imatinib. Out of the remaining nine CML patients treated with Nilotinib (300 mg), seven patients had optimal responses and the remaining two patients failed responses (Table 1). Overall, out of 21 CML treated patients with TKIs, 11 patients had optimal responses (52%), and ten patients had failed treatment (48%), (Table 1). In addition, the results of Fisher's Exact Test revealed that treatment with Nilotinib (300 mg) was more likely to lead to optimal response compared to treatment with Imatinib (400 mg) (Odds Ratio: 15.75, 95% CI 1.75-141.41, Z: 2.46, p < 0.05) ( Table 2A). Figure 1 shows the sequence of results and analysis that were performed using two techniques at different time points (diagnosis and follow up) for different comparisons.

Presentation of results.
Expression of BCR-ABL1 and PTPRG mRNA in whole white blood cells using RT-qPCR. The expression levels of BCR-ABL1 and PTPRG mRNA levels in CML patients at diagnosis and follow up were determined by RT-qPCR and the results are presented in Fig. 2. As the results of Cohn's d coefficient analysis show, there was a "huge" and "large" effect size on BCR-ABL1 (Cohen's d = 5.05) and PTPRG transcripts (Cohen's d = 0.81) following treatment with TKIs. A significant difference was found in the mean levels of BCR-ABL1 mRNA at diagnosis and at follow up (WSRT p ˂ 0.001, Fig. 2a) and PTPRG mRNA at diagnosis and at follow up (WSRT p ˂ 0.001, Fig. 2b). In contrast to BCR-ABL1 mRNA, which had significantly higher levels at diagnosis www.nature.com/scientificreports/ compared to follow up, PTPRG mRNA expression levels were found to be significantly lower at diagnosis compared to follow up. There was also a moderate negative correlation between BCR-ABL1 at diagnosis and PTPRG at follow up (r s (21) = − 0.422, p = 0.028). Moreover, the PTPRG transcript was also assessed to compare between optimal and failed CML groups. The results showed that the optimal response group had significantly higher PTPRG expression during follow up (WSRT p ˂ 0.0005) compared to diagnosis time-point, while no significant difference was found amongst the failed group at diagnosis and during follow up (WSRT p = 0.312, Fig. 2c).

PTPRG expression levels on white blood sub-populations (WBCs) of Healthy individuals and
CML patients at diagnosis using flow cytometry. Next, the expression levels of PTPRG protein were determined in the WBCs sub-populations of healthy and CML patients using flow cytometry and the results are presented in Figs. 3 and 4. A significant difference was found in the expression levels of PTPRG between different WBC sub-population (neutrophils, monocytes, and lymphocytes) [F (3, 5) = 19.15, p ˂ 0.0001]. PTPRG protein expression levels were found to be higher on neutrophils and monocytes when compared to lymphocytes in both healthy individuals (Fig. 3a,b) and CML patients at diagnosis (Fig. 3c).
When comparing healthy individuals against CML patients, PTPRG protein expression levels were significantly higher on neutrophils (U = 30, p < 0.002) and monocytes (U = 24, p < 0.007) amongst healthy individuals in comparison to CML patients at diagnosis (Fig. 3d). A significant difference was also found in the expression level of PTPRG on neutrophils between the healthy, optimal and failed groups (H = 14.94, df = 3, p = 0.001). The optimal response group had higher expression PTPRG in their neutrophils compared to the healthy group (U = 3, p = 0.004), and the failed groups had higher expression of PTPRG than the healthy group (U = 7, p = 0.004).
Expression levels of PTPRG protein on neutrophils and monocytes in optimal and failed groups at follow up determined by flow cytometry. The PTPRG expression level on WBCs was also re-assessed during the follow up for both optimal and failed CML groups. There was a significant difference between the expression of PTPRG on the neutrophils (χ 2 (2, 11) = 13.82, p = 0.001) and monocytes (χ 2 (2,   Fig. 5a). However, for the monocytes there were only significant differences in its expression levels between 1st follow up (median: 3.5) and 3rd follow up (median: 4.5, Z = − 2.05, p = 0.004, Fig. 5). There were no significant differences between the follow up time-points in relation to the expression of PTPRG on the lymphocytes in the optimal response group (χ 2 (2, 11) = 2, p = 0.6).
When PTPRG expression was compared on neutrophils of optimal and failed group, there was no significant difference in the 1st follow up (U = 39, p = 0.26). On the other hand, there were significant differences between the two response groups at the 2nd follow up (U = 25.5, p = 0.036) and the 3rd follow up (U = 28.0, p = 0.05) (Fig. 5c). No significant differences in the monocytes of the optimal and failed groups were found between the follow up time points (1st follow up (U = 46.0, p = 0.53), 2nd follow up (U = 49.5, p = 0.7) and 3rd follow up (U = 48.0, p = 0.62)) ( Fig. 5f).
In CML patients, the impact of treatment (Nilotinib vs. Imatinib) on neutrophils' PTPRG expression significantly differed over time (Nilotinib: χ 2 (3, 8) = 18.45, p = 0.001; Imatinib: (χ 2 (3, 12) = 32.9, p = 0.001). There and failed treatment groups. The mRNA level of PTPRG in the optimal response group was significantly higher when compared with diagnosis, while this significance was lost in the CML group. The Y-axis represents number of PTPRG/ABL1 of mRNA copies, while the X-axis represented timelines at diagnosis and mean of follow up. P values were derived from the WSRT test.   (Table 3). Furthermore, the mean rank of PTPRG expression on LSC at the time of diagnosis in the failed group was significantly higher than the optimal group (U = 12, p < 0.004).

Discussion
In this study, for the first time we investigated the changes in the PTPRG protein level in CML patients at the State of Qatar determined by flow cytometry using a unique monoclonal antibody against the external domain of PRPRG. We confirmed that restoration levels of PTPRG were significantly greater during the follow-up period following treatment with BCR-ABL1 TKIs compared to the time of diagnosis (Fig. 4b). Moreover, stratification of CML patients' response into optimal and failed treatment groups showed that the expression level of PTPRG to be significantly higher in the optimal response group when compared to the failed group, utilizing both q-PCR and flow cytometry techniques. Our results also confirmed our earlier findings using RQ PCR, in which 33 CML patients had been studied and the mRNA level of PTPRG was found to be significantly high in patient achieving Major Molecular Response (MMR), while this was not the case in non-responsive cases 17 .
There is currently no clear associations between the expression of a biomarker at mRNA level and its translation at protein level and the link between the two parameters ranged from 40% up to 90% [18][19][20] . On the contrary, other publications reported a low correlation between levels of mRNA and protein expression due to many factors including steady state, degradation, proteomics, and transcriptomics factors [21][22][23] . In the current study, we investigated the expression levels of PTPRG at both the mRNA and protein levels in parallel. While we did not find any significant differences in the expression of PTPRG at the mRNA level in the failed group (WSRT p = 0.312) (Fig. 2c), there were significant differences in the PTPRG protein expression level on neutrophils population in CML of the optimally treated group (Fig. 5c). Our results showed the importance of studying the expression levels of various biomarkers at both mRNA and protein levels.
In this study, we found that the restoration of the PTPRG level to be significantly higher on the neutrophils population in the optimal response group, when compared to the failed group at two follow up points (2nd and 3rd follow up) (Fig. 5c). On the other hand, there was not a significant difference in the expression of PTPRG on the monocytes population between the optimal and the failed groups at any follow up point (Fig. 5f). Our data also showed that the restoration of PTPRG expression on neutrophils was drug-dependent as the expression of PTPRG was restored earlier with Nilotinib when compared to Imatinib Mesylate. This observation might reflect the superior potency of Nilotinib as BCR-ABL1 TKI, which was reported to be 20-50 times more potent than Imatinib, and its ability to achieve complete superior response compared to Imatinib Mesylate 24 .
In another study, Naoto and his colleagues developed a fluorescence in situ hybridization (FISH) technique named "Neutrophil-FISH" that has the ability to classify CML cohort to responder and non-responder to Imatinib Mesylate 25 . In the current study, we reported the ability of mAB TPγ B9-2 to record changes in expression of PTPRG on neutrophils by flow cytometry technique and consequently classify the CML patients in the same manner.
The restoration of PTPRG expression reached the level recorded in healthy individuals and this may be explained by the recovery of healthy hematopoiesis in these subjects due to that TKIs having a large impact on the PTPRG gene (Cohen's d = 0.81). On the other hand, there was still a significant difference between levels of PTPRG in healthy, optimally, and failed groups. This may be explained by overexpression of PTPRG to overcome the uncontrolled BCR-ABL1 kinase activity.
In relation to expression of PTPRG on myeloid lineage, our previous study have shown a low level of PTPRG expression in CML on myeloid lineage at the time of diagnosis 17 . Interestingly, in this study, we found the mean PTPRG expression level to be higher in hematopoietic stem cells when compared to leukemic stem cells. The Glass' Delta effect size equation showed a small effect size of level of PTPRG at HSCs, as compared to a medium effect size of the level of PTPRG on LSCs (Table 3). Additionally, the expression of PTPRG on LSCs was significantly higher in the failed group when compared to the optimal response. The result was matched with the fact that leukemic stem cells have a unique cell surface profile, which is different from that of hematopoietic cells 26 . Furthermore, LSCs had the ability to self-renewal 27 , and LSCs had a signatory high expression of a gene, which is independently associated with adverse outcomes of treatment 28 and could predict the prognosis of the disease 29 . A recent study documented that BCR-ABL1 transcripts maybe not transcribed by LSC of CML patients 30 .
Finally, primary resistance and acquired resistance to treatment with BCR-ABL1 TKIs can occur in some patients with CML [31][32][33][34] . In Qatar, a significantly higher percentage of CML patients develop resistance to TKIs compared to the other parts of the world 12,35 . This may be due to the fact that most studies in CML management were focused on one component (tyrosine kinases), while the other arm (tyrosine phosphatases) has not yet received equivalent attention.
In summary, in this study, we have shown that the expression of PTPRG, a tumor suppressor gene, is suppressed in CML patients and this can be restored following treatment with TKIs to levels observed in healthy controls. We have also shown that restoration levels were greatest in optimal responders and occurred earlier with nilotinib compared with imatinib. Taken together, our results support that determination of PTPRG expression level by flow cytometry as a new biomarker of response to treatment with BCR-ABL1 TKIs is a useful tool for studying its role in tumor progression and predicting the response to therapeutic interventions and clinical management in patients with CML. Blood samples were collected from newly diagnosed patients (ND) at day zero and before starting any tyrosine kinase inhibitor (TKI) or Hydroxyurea treatment. For failed/relapsed patients' (F), blood samples were collected at the time of failure. The CML patients' response to treatment was assessed based on the hematologic, cytogenetic, and molecular responses according to the European LeukemiaNet (ELN; 2013). The absence of one of the following was considered as treatment failure: abnormal complete blood count and/or Ph + > 95% by 3 months, BCR-ABL1 > 10% and/or Ph + > 35% by 6 months, or BCR-ABL1 > 1% and/or Ph + > 0 by 12 months treatment 36 . The Optimal response group (R) was defined as achieving hematological, molecular, and cytogenetic remission within the time points of ELN guidelines.

Determination of PTPRG and BCR-ABL1 transcripts by RT-qPCR.
A RT-qPCR was carried out to measure the mean levels of mRNA of PTPRG and BCR-ABL1 at 3, 6 and 12 months. Plasmids carrying PTPRG full length were utilized to standardize the number of copies measured by PCR as previously described 9,17,35,37,38 . IPSOGEN BCR-ABL1 Mbcr IS-MMR Kits CE kit (Cat No. 670823) was utilized to quantify BCR-ABL1 according to the international scale (IS). RT-qPCR and Flow cytometry measurements were done simultaneously. Flow cytometry was performed on peripheral blood cells (PBCs) to characterize the expression of PTPRG on lymphocytes, monocytes, and neutrophils (LMN). The gating strategy of the flow cytometry experiment aimed at monitoring PTPRG expression in the major populations represented by neutrophils, monocytes and lymphocytes at diagnosis and during the follow up phase of TKIs treatment (Fig. 4a,b).  Neutrophils and Monocytes showed a low level of PTPRG expression at the time of diagnosis. PTPRG restored its expression, at least in part of the sub-population of white blood cells followed by TKIs therapy. Of note, lymphocytes remained at a low level acting as an internal control. Follow up time points F1, F2 and F3 were 3, 6, and 12 months of successful TKIs as per ELN timelines. The MFI values were obtained by calculating the ratio differences between the signals derived from the signal of mAB TPγ B9-2 and irrelevant mouse IgG1. (c) Gating strategy to identify leukemic CD34+CD38− stem cells. For myeloid progenitors and its sub-population, we targeted 15-20% upper and lower population of CD34 (red color) with CD38, with the upper population corresponding to the target of interest (hematopoietic stem cells). In comparison the lower population correspond to part of the leukemic stem cell.   www.nature.com/scientificreports/ a "huge" effect size 39,40 . Additionally, a Spearman's rank-order correlation was run to determine the relationship between the two genes. The distribution of the data was tested for normality with the Shapiro-Wilk test (W), and Skewness and normally distributed data was tested by the One-way Analysis Of Variance (ANOVA) and Dunnett's multiple comparison tests. Non-normally distributed data was tested by Friedman and Dunnett's multiple comparison tests. On the other hand, the Mann Whitney test was performed to analyze the difference between normally and non-normally distributed data. Changes in expression of PTPRG in CML patients over timeline were analyzed with Friedman and Wilcoxon signed ranks (WSRT) tests and a pairwise comparison with Hodges-Lehmann estimator. Finally, Kruskal-Wallis, along with Mann Whitney tests, were employed to examine the changes of PTPRG among healthy (H), optimally (R), and failed (F) groups. Consent for publication. Consent for publication was obtained through ethics approval and consent to participate.

Data availability
This is a research article, and all data generated or analyzed during this study are included in this published article. Please contact the corresponding author for data and material requests. Monoclonal Antibody for Diagnosing Chronic Myeloid Leukemia Patent Hamad Medical Corporation sponsored the patent process of the antibody and assignee: Qatar foundation. Provisional US patent application number: US62373322.