In clinics, a sensitive, quick, convenient and versatile detection method to measure aberrant kinase activity is desirable for many pathologies, including various forms of leukemias, and could be important to improve diagnosis and treatment. The Abelson (ABL)-1 tyrosine kinase (TK) belongs to the non-receptor TK family and is an ubiquitously expressed cytosolic enzyme that plays a role in many key processes linked to cell growth and survival as well as to cell differentiation, cell adhesion and migration (Greuber et al., 2013). Oncogenic forms of ABL1, in particular the BCR-ABL1 fusion gene with different breakpoints in BCR gene, are well-recognized for their pathogenic role in leukemias, including chronic myeloid leukemia (CML), some forms of acute myeloid leukemia (AML), and acute lymphoblastic leukemia (ALL). BCR-ABL1 results from the chromosomal translocation t (9;22), and encodes for the chimeric BCR-ABL1 protein that leads to an aberrant costitutive activation of the ABL1 proliferation pathway. In the chimeric BCR-ABL1, the N-terminal BCR region interferes with the negative regulation of ABL1 and mediates dimerization of the protein that autophsophorylates the kinase, fully and constitutively activating it (Greuber et al., 2013). Moreover, in ALL, BCR-ABL1 like forms exist [10–15% of B-ALL pediatric cases, incidence increased with age (Roberts et al., 2014a)], presenting blasts that are transcriptionally related to BCR-ABL1 expressing cells althought they lack the t(9;22) transloctation. These forms are recognized as a high-risk ALL subtype across all clinical studies, being characterized by an unfavorable prognosis and a higher relapse rate in comparison to other B-ALL subtypes (Roberts et al., 2014a; Boer et al., 2015). BCR-ABL1 like cells harbour a multitude of other genomic rearrangements that involves either the ABL encoding genes (e.g., ABL1, ABL2), the JAK-kinase genes (e.g., JAK1, JAK2) or their upstream receptors (e.g., PDGFRB, CSF1R. CRLF2, EPOR) (Roberts et al., 2012; Boer et al., 2017), resulting in the presence of chimeric TK costitutively activating either the ABL1 or the JAK-STAT proliferation pathway (Roberts et al., 2014b; Boer and den Boer, 2017). A third, smaller group is characterized by involvement of genetic fusions that lead to altered TK in the RAS pathways (Ofran and Izraeli, 2017).

In reference to the development of a peptide biosensor to detect ABL1 activity, it is important to consider that ABL1 mediates its catalytic function by protein-protein interactions through three SRC homology domains (highly conserved SH3–SH2–SH1 cassette) located at the ABL1 N-terminal. SH1 binds and cleaves ATP, and mediates tyrosine (Y) phosphorylation in target substrate proteins; SH3 and SH2 domains function as interaction modules for targets and as allosteric inhibitors of the catalytic SH1 domain (Corbi-Verge et al., 2013). The consensus motif of an ABL1 substrate is I/V/-Y-X-X-P/F (where X is any amino acid, Y representing phosphorylation site). This sequence has been considered as the base to develop a peptide biosensor to detect ABL1 activity and was found by screening a library of billion distinct peptides for interaction with purified ABL1 (Songyang et al., 1994; Songyang et al., 1995). Phospho-peptides generated by the incubation with the TK of interest were isolated and sequenced to determine the optimal aminoacid tolerated at each position. The optimized aminoacidic sequence EAIYAAPFAKK (named as “reporter” sequence in this manuscript), was designed. This sequence represents an artificial kinase substrate, not a naturally occurring motif of endogenous proteins (Songyang et al., 1994; Songyang et al., 1995) and corresponds to the commercially available Abltide (P_ABLTIDE), nowadays largely accepted as a reference peptide biosensor for screening ABL1 activity. However, P_ABLTIDE is suitable only for biochemical research use (i.e., enzymatic assays on purified ABL1 or pathogenic chimeric ABL1 proteins), and is not intended for use in clinics and diagnostics because of the lack of serum stability, of cell-penetrating properties and of specific ABL1 targeting among other kinases in primary cells (Wu et al., 2010; Henriques et al., 2013). An additional optimized stretch of amino acids (APTYSPPPPP, named as “targeting” in this manuscript) has been designed using biocomputing tools on the basis of physicochemical reasoning, and is recognized by the SH3 domain of ABL1 protein (Pisabarro and Serrano, 1996). This sequence could be linked to the reporter region, resulting in a novel artificial peptide (named P_ABL in this manuscript) with an increased specificity for ABL1. This peptide was further developed to detect intracellular ABL1 kinase activity in live intact cells (Placzek et al., 2010; Tang et al., 2012; Yang et al., 2013), and was applied to the investigation of the activity of the chimeric BCR-ABL1 in whole cell lysates of patients affected by CML (Yang et al., 2013).

The introduction of targeted therapy with imatinib and other small-molecule inhibitors that target BCR–ABL1, competing at the level of the ATP binding site of the kinase, has significantly improved the outcome in patients with BCR-ABL1 positive leukemias (An et al., 2010; Biondi et al., 2018). It was thus hypothesized that TKIs specific to ABL1 (e.g., imatinib) or JAK (e.g., ruxolitinib) might be valuable options also for BCR-ABL1 like ALL cases of the ABL-class or JAK-class respectively (Boer et al., 2017; Fazio et al., 2020), and might represent a better therapeutic choice for these patients, who respond poorly to conventional chemotherapy (Den Boer et al., 2009; Roberts et al., 2012; Boer et al., 2015). Nowadays, several ongoing studies are assessing this issue (clinicaltrials.gov identifiers: NCT02883049, NCT02723994, NCT03117751, NCT02420717, NCT03571321). The heterogeneity in their genetic background makes the identification of BCR-ABL1 like ALL very challenging. Currently, the molecular analyses carried out on patient’s blasts at diagnosis include karyotyping, fluorescence in situ hybridization (FISH) or RT-PCR to detect few specific rearrangements. Useful high-throughput genomic approaches such as exome, whole genome and whole transcriptome sequencing are still too expensive to be routinely employed in clinical practice on large patients’ cohorts, particularly in low-income countries (Boer et al., 2017). However, none of these genetic methods can give functional information about the aberrant kinase activity or about blasts sensitivity to TKI. This highlights the importance of having functional assays for TKs with key pathogenic role rather than genomic characterization as tools of precision therapy in BCR-ABL1 like ALL patients (Franca et al., 2018). Functional TK assays would represent an important tool for clinicians also for BCR-ABL1 CML and ALL: they could be used to guide the choice of TKI for the best target therapy and to identify primary TKI resistance in a timely manner at diagnosis. Indeed, both imatinib and dasatinib pose resistance problems in BCR-ABL1 positive patients: ABL1 kinase domain mutations have been detected in 30–90% patients who failed imatinib and in 20–80% of patients who failed dasatinib (Gunnar Cario et al., 2020). Identifying in a timely manner the best TKI to be used at diagnosis or during treatment would optimize the chances of survival in patients, and would overcome the possible issue of primary and acquired TKI resistance.

The aim of this study was to investigate the role of P_ABL (“reporter” + “targeting” sequence) as a biosensor able to quantify the ABL1 phosphorylation activity in whole lysates of immortalized leukemic cell lines, including BCR-ABL1 like ALL cells. Kinase activity was monitored in vitro by an ELISA assay. Long-term goal is to set up a point-of-care device to improve the diagnosis and clinical management of BCR-ABL1 and BCR-ABL1 like leukemias in clinics. Far from overcoming the importance of monitoring the minimal residual disease, these devices could nonetheless boost the therapeutic monitoring by measuring the residual TK activity or the acquired resistance to these drugs in residual blasts.

In recent years, several potential peptides allowing the study of kinases’ activity in vitro have been developed (Lipchik et al., 2015; Perez et al., 2019). The artificial peptide P_ABL resembles the previously published cell-permeable biosensor peptide used to monitor the BCR-ABL1 function in CML (Yang et al., 2013), with “reporter” and “target” sequences in common. Here, P_ABL was employed to validate its use in lysates of ALL cell lines with different genetic background, including those harboring BCR-ABL1 like ALL translocations of the ABL-class. The phosphorylation of the P_ABL probe was detected after the incubation with cell lysates. Interestingly, none of the cell lysates induced an increase of P_ABL-F phosphorylation, indicating a kinase discriminatory action on the tyrosine in the “reporter” region. K562 cells showed the highest fluorescence signal on P_ABL although comparable to that observed for ALL-SIL. Both these cell lines strongly rely on ABL1 for survival and proliferation, as demonstrated by MTT and trypan blue viability assays, although only BCR-ABL1 and not NUP214-ABL1 could be detected in the phosphorylated form by western blot. It is known that NUP214-ABL1 displays much lower auto-phosphorylation than BCR-ABL1 (Brasher and Van Etten, 2000; De Keersmaecker et al., 2008a; De Keersmaecker et al., 2008b). Because of this lower activation, ALL-SIL cells are more sensitive than K562 to imatinib, an ABL1 inhibitor that targets specifically the inactive conformation of the ABL1 kinase. Nonetheless, BCR-ABL1 and NUP214-ABL1 largely overlap in their substrate specificity as demonstrated using a peptide array (De Keersmaecker et al., 2008a).

NALM6 cells were chosen because they harbor the gene fusion ETV6-PDGFRB. PDRGFB is a cell surface TK receptor that activates the ABL1 pathway (Plattner et al., 2004), and PDRGFB rearrangements (ETV6-PDGRFB and EBF1-PDGFRB) are present in about 1% of BCR-ABL1 like positive patients (Boer et al., 2017). Recent studies observed a reduction in signalling with imatinib in cases harboring ABL1-class rearrangments (Roberts et al., 2017), and a case-report on a refractory EBF1-PDGFRB ALL patient showed the complete remission after the addition of imatinib to the conventional chemiotherapy (Weston et al., 2013). The chimeric protein ETV6-PDGFRB contains the amino-terminal 154 amino acids of ETV6 fused to the transmembrane and cytoplasmic domains of PDGFRB, and has a calculated molecular mass of 76 kDa (Carroll et al., 1996). Western blot using an anti human PDGFRB confirmed the presence of ETV6-PDGFRB in NALM6: the monomer migrated at 90–100 kDa in SDS gel-electrophoresis; however, ETV6-PDGFRB appeared in NALM6 also as oligomeric or multimeric complexes resulting in a band around 200 kDa, as already observed in literature (Sjöblom et al., 1999). The wild type PDGFRB was also detectable in all the cell lines (a band around 100 kDa) except for K562. Some investigators report that NALM6 cell line harbors a different translocation, in particular DUX4-rearranged [t (4;14) (q35;q32)]/ERG deletion (Yasuda et al., 2016; Tanaka et al., 2018); however, both our western blotting analysis and cytogenetic information on the t(5;12)(q33.2;p13.2) translocation confirm the presence of the ETV6-PDGFRB in the NALM6 cells used in this paper (Wlodarska et al., 1997) ¹ . Since PDGRFB is linked to ABL1 activation (Plattner et al., 2003), we expected higher P_ABL phosphorylation values also for the BCR-ABL1 like cell line NALM6, in particular in comparison to REH cells that should not present constitutively active ABL1 based on their genetic profile, but surprisingly, these two cell lines showed similar fluorescence levels. The reason is still unclear, but can be related to the ubiquitous presence of ABL1 in the cell lines (Greuber et al., 2013). Indeed, Western blot analysis with anti-human ABL1 showed the presence of non-phosphorylated ABL1 in both REH and NALM6, without detecting the phosphorylated form.

The specificity of the P_ABL phosphorylation signal was also confirmed after treatment with TKI. Indeed, the ABL1-inhibitor imatinib was used at a concentration of 5 μM that, as reported in the literature, represents the steady state plasma concentration after 5–7 days of treatment in adult CML at a dose of 400 mg/day (Druker et al., 2001), and is 20 times higher than the IC₅₀ calculated in vitro on cellular tyrosine phosphorylation assay (Deininger et al., 2005). At this concentration, a significant decrease in P_ABL phosphorylation levels after the incubation with the drug was observed for all cell lines. It is known from the literature that imatinib has an unusually high selectivity to ABL1 because it targets the inactive conformation which is unique to this kinase (Lee and Wang, 2009). Mass spectrometry analysis investigating imatinib-associated proteins in K562 lysates confirmed the direct interaction of the drug with only few targets, including the known interactor BCR-ABL1 and the ABL-releted gene (ARG), Discoidin Domain Receptor Tyrosine Kinase 1 (DDR1) and the proto-oncogene receptor tyrosine kinase (KIT) (Bantscheff et al., 2007; Rix et al., 2007). DDR1 and KIT are receptor tyrosine kinases and their contribution to the ABL1 activation likely depends on extracellular stimuli. Therefore, considering that our in vitro system works on cell lysates rather than whole cells in the absence of specific ligands, their confounding contribution to the P_ABL phosphorylation, if any, is likely less important. In contrast, contribution of ARG (closely related to ABL1) could be an issue. However, the P_ABL target region should guarantee an increased specificity for ABL1 compared to ARG. Mass spectrometry studies identified additionally one off-target (non-tyrosine kinase) of imatinib, the NAD(P)H:quinone oxidoreductase NQO2, whose enzymatic function should not influence the peptide phosphorylation (Bantscheff et al., 2007; Rix et al., 2007). In contrast to imatinib, the anti-JAK inhibitor ruxolitinib did not affect the P_ABL phosphorylation levels at any of the concentrations selected (52 nM and 5 μM). The concentration of 52 nM was chosen in accordance to pharmacokinetic parameters measured in healthy subjects at steady state following a twice-daily administration of ruxolitinib at 15 mg and is 20 times higher than the IC₅₀ of 2.8 ± 1.2 nM calculated in vitro for JAK2 by a biochemical enzymatic inhibitory assay (Quintás-Cardama et al., 2010). The second concentration of 5 μM is much higher than the pharmacological range and was included into the P_ABL-based ELISA assay in order to verify the absence of inhibition, even at this high concentration.

In our in vitro ELISA assay, P_ABLTIDE was clearly less performing than P_ABL, designed to increase the specificity to ABL1 among other kinases. Besides the increased specificity of P_ABL over P_ABLTIDE, the ELISA assay proposed in this paper presents other advantages over conventional kinase assay. Firstly, the biosensor phosphorylation was quantified on a solid-phase through a neutravidin-coated plate that allows the binding of the biotin in the P_ABL sequence. This high specific binding greatly reduces background noise signals. Secondly, whole cell lysates can be used, avoiding tricky ABL1 purification and bypassing the problems related to the peptide penetration into intact cells. Thirdly, a unique primary antibody, i.e., an anti-phosphotyrosine, is required. This advantage is particularly important for BCR-ABL1 like ALL because the ELISA assay conditions could be optimized regardless of the specific chimeric protein encoded by the genetic abnormalities in BCR-ABL like leukemic cells of the ABL-class.

Initial attempts to validate the novel P_ABL-based ELISA assay on lysates of patients’ leukemic cells were performed. Although results are preliminary and limited to only two samples, they allow some considerations, to be confirmed by further investigations. As observed in immortalized cell lines, P_ABL becomes phosphorylated after the incubation with leukemic cell lysates regardless the presence of a BCR-ABL1 chimeric protein; however, in contrast to them, this phosphorylation is ABL-1 mediated only in some patients in which it is specifically inhibited by imatinib. A different contribution of ABL1 and other TK on P_ABL could be hypothesized among different patients, according to their pathogenetic profile.

Taken together, these observations suggest that the P_ABL-based ELISA assay is suitable for measuring the ABL1 kinase activity of cell lysates through the P_ABL-Y phosphorylation in the “reporter” region. P_ABL could be more suitable in detecting aberrant sustained kinase activity of ABL1-chimeric proteins rather than an over-activation of native ABL1 due to upstream signaling; however, a limitation of this study is that the difference between basal and sustained ABL1 activity is still too tiny for the P_ABL-based ELISA assay to be of practical clinical interest without any further optimization. To confirm its utility, our system should be improved in sensitivity and accuracy, and results could be strengthened by other TKI, the use of different biosensors (e.g., of downstream signaling components such as STAT5 that is known to be activated by ABL1-class rearrangements), and the test of patient samples of known genotype. Nonetheless, the results here described represents the first step towards the setup of a point-of-care device for diagnosis and therapeutic drug monitoring of pediatric BCR-ABL1 like patients.