The EphB6 receptor is overexpressed in pediatric T cell acute lymphoblastic leukemia and increases its sensitivity to doxorubicin treatment

While impressive improvements have been achieved in T-ALL therapy, current treatment approaches fail in approximately 25% of patients and these patients have limited treatment options. Another significant group of patients is being overtreated, which causes long-lasting side effects. Identification of molecules controlling drug resistance in T-ALL is crucial for treatment optimisation in both scenarios. We report here the EphB6 receptor is frequently overexpressed in T-ALL. Remarkably, our observations indicate that EphB6 acts in T-ALL cells to enhance sensitivity to a DNA-damaging drug, doxorubicin, as interruption of EphB6 activity interferes with the efficiency of doxorubicin-induced eradication of T-ALL cells in cell culture and in xenograft animals. This effect relies on the protection of Akt kinase signaling, while Akt inhibition combined with doxorubicin application produces synergistic effects on the elimination of EphB6-deficient T-ALL cells. These data imply that EphB6 suppresses T-ALL resistance by interfering with Akt activity. Our observations highlight a novel role for EphB6 in reducing drug resistance of T-ALL and suggest that doxorubicin treatment should produce better results if personalised based on EphB6 levels. If successfully verified in clinical studies, this approach should improve outcomes for T-ALL patients resistant to current therapies and for patients, who are being overtreated.

patients. Therefore, there is an urgent need to identify molecules that determine drug resistance of T-ALL cells, as this knowledge should assist in improving existing treatment approaches.
The Eph group of receptor tyrosine kinases is presented by 16 Eph receptors, of which 14 are expressed in human cells [13][14][15] . Through their basal or ligand-induced action, these molecules control a vast variety of signaling events and cellular responses in both normal, and malignant cells 16 . Signaling activity of Eph receptors is modulated by their ligands, ephrins, where Eph receptors interact in a very promiscuous manner with ephrin-A (ephrin-A1 -ephrin-A6) or ephrin-B (ephrin-B1 -ephrinB3) types of ligands 17 . Based on their structural properties and ligand binding preferences, Eph receptors are divided into EphA or EphB subgroups, with EphA receptors binding mostly ephrin-A and EphB receptors interacting predominantly with ephrin-B molecules. In a classical model, ephrin binding induces Eph receptor dimerisation or oligomerization, leading to the phosphorylation on tyrosine residues, which further enhances catalytic activity of Eph receptors and supports their interaction with cytoplasmic signaling partners 13,16 .
Interestingly, two members of the Eph group, EphA10 and EphB6, are devoid of kinase activity because of several innate alterations in their kinase domains 15,[18][19][20] . Accumulating evidence suggest that despite the lack of kinase activity, EphB6 undergoes tyrosine phosphorylation that can be provided by some catalytically active Eph receptors 21,22 or Src family kinases 23 and has important functions in T-lymphocytes [24][25][26] . Moreover, our previous work shows that along with the some other EphB receptors, EphB6 is also expressed in the majority of the analysed T-ALL cell lines and patient samples, where a collective action of EphB receptors protects T-ALL cells from Fas-induced apoptotic death 27 .
In our investigation discussed in this manuscript, we examined the role of EphB6 in T-ALL drug resistance. Our new findings reveal that EphB6 is the only member within the Eph group that is overexpressed in the majority of T-ALL cases. Remarkably, our observations also indicate that in contrast to the collective action of EphB receptors, EphB6 does not support survival of T-ALL cells, but rather increases their sensitivity both in cell culture and in vivo to a DNA-damaging therapeutic compound, doxorubicin, that is being commonly used for T-All treatment 11,28 . On a molecular level, this effect is associated with the decrease in the activating phosphorylation of a pro-survival molecule, the Akt kinase and with reduced phosphorylation of its downstream target, the p70 S6 kinase, indicating that EphB6 acts by suppressing Akt signaling in doxorubicin-treated T-ALL cells.
Taken together, our work identifies a previously undescribed function for the EphB6 receptor in controlling drug sensitivity of T-ALL cells and suggests that the efficiency of DNA-damaging therapeutic reagents, including doxorubicin, could be improved by applying them in a personalized manner to patients with different levels of EphB6 expression in T-ALL cells.

Results
The EphB6 receptor is overexpressed in the majority of T-ALL cases. Previously published observations, including the work of our team demonstrate that expression of the EphB6 receptor is maintained in most analysed T-ALL cell lines and some patient samples 27,29 . However, EphB6 expression, and expression of other Eph receptors and their ligands, ephrins, has never been systematically assessed in a larger collection of T-ALL cases in comparison to matching normal controls. To address this, we used the dataset from the European Bioinformatics Institute and analysed expression of ephrins and Eph receptors in 117 pediatric T-ALL patient samples, comparing it with matching normal tissue samples. The statistical significance of this analysis was verified using the Mann-Whitney U test. Excitingly, this approach revealed that EphB6 is the only molecule within the Eph/ ephrin group, whose expression is significantly elevated in the majority of T-ALL cases (Fig. 1). Moreover, the Figure 1. EPHB6 is overexpressed in T-ALL patient samples. Eph receptor and ephrin expression was analyzed in 117 pediatric T-ALL patient samples and 7 matching normal tissue controls using data from the European Bioinformatics Institute. Statistical significance was computed using the Mann-Whitney U test. Expression of the Notch3 receptor that is upregulated in T-ALL is shown as a reference.
Scientific RepoRts | 7: 14767 | DOI:10.1038/s41598-017-15200-3 quantification of our bioinformatics data also showed that EphB6 is overexpressed by at least two fold in ~76% of T-ALL cases, indicating that the EphB6 receptor may produce a prominent effect on the biological properties of T-ALL cells.

EphB6 suppresses drug resistance of T-ALL cells.
Our previous investigation showed that the collective action of EphB receptors inhibits Fas-induced apoptotic response in T-ALL cells 27 . Since we found that EphB6 is overexpressed in at least 76% of T-ALL patients, we suspected that it may also suppress apoptotic responses induced by chemotherapeutic drugs. To asses this possibility, we used in our experiments a previously described stable cell line originating from human T-ALL cells Jurkat, ΔB6-Jurkat, that had been transfected with a dominant-negative EphB6 mutant (DN-EphB6) 24 . This mutant maintains intact extracellular and transmembrane domains, but lacks the cytoplasmic portion and interferes with EphB6-mediated responses 24 . Jurkat cells mock-transfected with the empty pcDNA3 expression vector (pc3-Jurkat) were used as a specificity control ( Fig. 2A). Contrary to our expectations, ΔB6-Jurkat proved to be significantly more resistant to the treatment with a DNA-damaging therapeutic compound, doxorubicin, which is being used in T-ALL therapy 11,28 , than control pc3-Jurkat cells (Fig. 2B). To confirm the specificity of this observation, we partially silenced EphB6 expression in T-ALL cells, E6.1, that have the same origin as Jurkat (Fig. 2C). Consistent with the effect of the DN-EphB6 mutant, EphB6 silencing also reduced sensitivity of T-ALL cells to doxorubicin (Fig. 2D). In these experiments, EphB6-targeting shRNA produced a relatively milder effect on increasing drug resistance, which was most probably due to the incomplete silencing of EphB6 expression (Fig. 2C). Taken together, these data indicated that EphB6 acts differently from other EphB receptors that collectively suppress apoptotic cell death in T-ALL cells and instead, enhances T-ALL sensitivity to doxorubicin treatment. This effect proved to be not cell line-specific, as expression of DN-EphB6 in human malignant lymphoblastic T cells, H9, also increased their resistance to doxorubicin ( Fig. 2E and F).

EphB6 action in increasing drug sensitivity of T-ALL cells is associated with enhanced apoptotic response and decreased Akt activation.
To determine if EphB6 increases drug sensitivity in T-ALL by supporting the induction of the apoptotic cell death, we monitored activation status of a key pro-apoptotic molecule, caspase-3 30 , in ΔB6-Jurkat and pc3-Jurkat cells using the EnzChek Caspase-3 assay. This approach revealed that inhibition of EphB6 activity strongly inhibits caspase-3 activation in doxorubicin-treated T-ALL cells ( Fig. 3A and B). Consistent with this, expression of DN-EphB6 also blocked doxorubicin-induced cleavage of a caspase-3 substrate, poly (ADP-ribose) polymerase (PARP) 31 (Fig. 3C). Moreover, our flow cytometry analysis of annexin V staining demonstrated that suppression of EphB6 action by the dominant negative mutant or shRNA-based silencing efficiently reduces doxorubicin-triggered apoptotic response (Fig. 3D).
Induction of the apoptotic cell death is negatively regulated through the activation of the Akt kinase and therefore, we assessed activating phosphorylation of this molecule on Ser473 residue by Western blotting with a phospho-specific antibody. In line with the pro-apoptotic action of EphB6, interruption of its signaling was associated with the lack of decrease in Akt phosphorylation in doxorubicin-treated cells (Fig. 4A). This was also accompanied by the maintained level of phosphorylation of a downstream target of the Akt-initiated signaling cascade, p70 S6 kinase 32 (Fig. 4B). The EphB6 receptor appeared to affect Akt signaling in a selective manner, as no effect of DN-EphB6 on activating phosphorylation of Erk kinases that also support cell survival 33 could be observed in doxorubicin-treated cells (Fig. 4C). Consistent with the model, where disruption of EphB6 signaling rescues T-ALL cells by protecting Akt activation, co-administration of doxorubicin and Akt inhibitor, perifosine, synergistically enhanced elimination of ΔB6-Jurkat cells (Fig. 4D).
Overall, these observations strongly suggest that EphB6 increases drug sensitivity by suppressing Akt signaling in doxorubicin-treated T-ALL cells, which ultimately enhances the apoptotic response.

Interruption of EphB6 signaling increases drug resistance of T-ALL cells in vivo. The ultimate
test for the drug sensitivity of cancer cells should be performed in an in vivo model. To achieve this, we followed previously described protocols 34,35 and injected ΔB6-Jurkat and pc3-Jurkat cells pre-mixed with matrigel into the flanks of immunodificient NOD.Cg-Prkdc scid Il2rg tm1Wjl /SzJ (NOD-SCID) mice. Resulting tumours were allowed to develop to a measurable size and treated with intravenous tail injections of doxorubicin. While expression of DN-EphB6 has not affected tumour growth (Fig. 5A), it dramatically decreased the effectiveness of doxorubicin treatment ( Fig. 5B-E). These in vivo data matched our initial observations obtained in cultured cells and provided a strong support for a model, whereby EphB6 action suppresses T-ALL resistance to treatment with DNA-damaging reagents.

Discussion
Acute lymphoblastic leukemia (ALL) is the most common pediatric malignancy and is the main cause of disease-related death in children. T-ALL is an aggressive ALL subtype that constitutes approximately 15% of pediatric ALL cases. During the last several decades, gradual fine-tuning of combination chemotherapeutic approaches dramatically improved T-ALL treatment and patient survival. Nevertheless, current treatment protocols fail in approximately 25% of T-ALL patients, and these patients cannot be cured 2,36 . Moreover, approximately 25% of children with T-ALL responding to chemotherapy are currently being overtreated, when unnecessarily aggressive therapy is being applied 12 . This over-aggressive treatment is often associated with severe long-lasting adverse side effects. All these deficiencies of current treatment protocols clearly indicate a strong need for more personalised therapeutic approaches and for less toxic targeted therapies. Identification and characterisation of molecular mechanisms that control responsiveness of T-ALL cells to chemotherapeutic reagents is crucial for treatment personalization, and for the development of effective targeted therapeutic strategies. In our work presented here, we found that a kinase-inactive member of the Eph group of receptor tyrosine kinases is the only Eph receptor that is significantly overexpressed in the majority of T-ALL cells. Since very high levels of EphB6 expression were previously observed in thymus and thymocytes 19,29 , it is likely that EphB6 overexpression does not directly contribute to T-ALL development, but rather represents a bystander effect of thymocyte transformation and reflects immature phenotypes of T-ALL cells. This matches our observations, showing that EphB6 does not accelerate propagation of T-ALL cells in a xenograft model and instead, reduces aggressiveness of T-ALL cells by making them more sensitive to the treatment with a DNA-damaging chemotherapeutic drug, doxorubicin. Obviously at this stage, we cannot completely exclude a possibility that EphB6 overexpression supports malignant transformation in the T cell lineage and this possibility is currently under an active investigation in our laboratory. Our previous work showed that EphB receptors collectively act through enhancing Akt activation to The graph shows caspase-3 activity in arbitrary units (AU), as measured at A 442 using a SpectraMax M5 plate reader, and represents analysis of one of three independent experiments. (B) The graph shows fold-change in caspase-3 activity following Dox treatment, compared to a matching solvent control and represents analysis of three independent experiments. (C) ΔB6-Jurkat and pc3-Jurkat cells were treated with 5 µM Dox for 14 hours at 37 °C and 5% CO 2 . Cells were lysed and cell lysates were resolved by SDS PAGE. PARP cleavage was analysed by Western blotting with anti-PARP. Western blotting with anti-tubulin was used as a loading control. To optimize presentation, anti-PARP and anti-tubulin Western blot images are shown at different brightness and contrast settings. (D) The indicated cells were treated with 5 µM Dox, or a matching solvent control, for 24 hours. Cells were stained with PE-conjugated Annexin V for 15 minutes in the dark at room temperature and analyzed by flow cytometry. The graphs on the right show median intensities of Annexin V staining of Doxtreated cells after the subtraction of values corresponding to median staining intensities of matching solventtreated controls. The graphs represent analyses of three independent experiments performed in duplicates. The analyses were done using the FlowJo software. All experiments were done at least three times. Full-length unadjusted Western blot images for this figure are shown in Supplementary Fig. 2A. *, P < 0.05, Student's t-test. n.s., statistically not significant. well, in 6 well plates) were treated with 5 µM Dox in serum-free medium for 14 hours. Cell lysates were resolved by SDS PAGE and activating phosphorylation of the Akt kinase was examined by Western blotting with anti-phospho-Akt, recognising Akt phosphorylation at Ser473 (anti-p-Akt). Western blotting with anti-GAPDH was used as a loading control. Akt phosphorylation was quantified by densitometry using Carestream software and measurements were normalized to matching GAPDH controls. The graph represents a reduction in p-Akt signal intensity in Dox-treated cells as a percentage relative to the matching solvent controls. To optimize presentation, anti-p-Akt and anti-GAPDH images are shown at different brightness and contrast settings. (B) Cells were treated as in (A) and phosphorylation of the p70 S6 kinase was analysed by Western blotting with anti-phospho-S6K (anti-p-S6K). Phosphorylation of p70-S6K was quantified and presented as in (A). To optimize presentation, anti-p-S6K and anti-GAPDH images are shown at different brightness and contrast settings. (C) Cells were treated as in (A) and activating phosphorylation Erk kinases was analysed by Western blotting with anti-phospho-Erk (anti-p-Erk1/2). Western blotting with anti-tubulin used as a loading control. To optimize presentation, anti-p-Erk1/2 and anti-tubulin images are shown at different brightness and contrast settings. (D) ΔB6-Jurkat (4 × 10 4 per well, in 96 well plates) were treated in triplicates with the indicated concentrations of Dox and with a suboptimal concentration of Perifosine individually, or in combination for 24 hours at 37 °C and 5% CO 2 . Cells were stained with resazurin and cell survival was monitored using a SpectraMax M5 plate reader. Data are shown as a percentage of cell survival in drug-treated populations relative to matching solvent controls. The synergistic effect of Dox/ Perifosine combinations was calculated using CompuSyn software and is presented as combination indices for each drug combination. Indices below 0.9 indicate synergism. All experiments were done at least three times. Full-length unadjusted Western blot images for this figure are shown in Supplementary Fig. 3A-C. *, P < 0.05, Student's t-test. n.s., statistically not significant.
suppress Fas-induced apoptosis and support survival of T-ALL cells 27 . Interestingly, our current data reveals that EphB6 has a completely opposite function in T-ALL cells, since inhibition of its action through the expression of a dominant-negative mutant or by shRNA-based silencing inhibits doxorubicin-triggered apoptotic cell death and this effect is supported by the increased Akt activation. This is consistent with our earlier observations, showing that EphB6 interacts with at least some kinase-active EphB receptors 21,22 and switches their pro-malignant action in cancer cells into anti-malignant responses 22 . Potential clinical relevance of our findings is further supported by in vivo experiments, confirming that EphB6 enhances doxorubicin-induced eradication of T-ALL cells not only in cell culture, but also in animal models. At this stage, our findings deliver an important message, suggesting that T-ALL treatment with DNA-damaging chemotherapeutic compounds could be efficiently personalised based on the level of EphB6 expression in T-ALL cells. Our observations suggest that patients with lower EphB6 levels could benefit from higher doses of doxorubicin, while lower doses could be sufficient for treating patients with high EphB6 expression. This approach could help to overcome resistance in some T-ALL cases, while also reducing undesirable side effects in patients, who are currently being overtreated. A careful matching of required were injected in a total volume of 100 µL subcutaneously into the right flank of NOD-SCID mice. As tumours reached a measurable size, mice were treated weekly with intravenous injections of Dox (2 mg/kg) or a matching volume of saline control (n = 7 per group). Tumour growth was monitored twice a week by measurements with digital callipers and tumour volume was calculated by the equation: A/2*B 2 , where A was long and B was short diameter of the tumour (A). The reduction in tumour growth in Dox-treated mice is presented as a percentage relative to matching saline-treated controls (B). In panels (A and B), day 0 indicates the day, when both treatment and measurements were initiated. Upon experiment termination, tumours were extracted, fixed in 10% formalin, photographed (C) and weighed. The average tumour weight for each experimental group is shown (D). The graph in panel (E) represents average tumour weights in Dox-treated mice as a percentage relative to matching saline-treated controls. Data are shown as means ± SD. One of two independent experiments is shown. *P < 0.05; Student's t-test. n.s., statistically not significant. doxorubicin doses to EphB6 levels in patients' T-ALL cells would be required in future for developing personalised treatment protocols.

Materials and Methods
Antibodies and reagents. Anti-EphB6 was obtained from Santa Cruz (Dallas, TX, USA). Anti-phospho-Akt (Ser473), anti-phospho-S6K, anti-phospho-ERK1/2, and anti-GAPDH were purchased from Cell Signaling Technology (New England Biolabs Ltd., Whitby, ON, Canada). Anti-PARP was purchased from EMD Millipore (Billerica, Mass, USA). Anti-β-tubulin was obtained from Santa Cruz Biotechnology (Dallas, TX . After 16 hours, medium was replaced with regular culture medium. Collected viral particles (after 48 and 72 hours of transfection) were concentrated using virus concentrator Lenti-X concentrator (Clontech Laboratories, Inc. CA, USA) according to the manufacturer's instructions. Cells were transduced by lentiviral particles in medium containing 10% FBS and 5 μg/mL polybrene (Sigma-Aldrich) in 24-well plates. Cells were incubated for 10-20 minutes at room temperature then plates were centrifuged at 1500 × g for 1 hour and 30 minutes at 35 °C. After centrifugation, cells were incubated at 37 °C and 5% CO 2 overnight. The lentiviral transduction medium was replaced with a complete culture medium. Cell lines with stable shRNA expression were generated by puromycin selection at 4 μg/mL for 3-4 days. Western blotting was performed to confirm EphB6 silencing.
Flow cytometry. The indicated T-ALL cells (1 × 10 6 per well) were incubated with 5 μM doxorubicin for 24 hours. Cells were harvested and washed twice with PBS. Pellets were resuspended in 250 μL 1X binding buffer from PE Annexin V Apoptosis Detection Kit 1 (BD Biosciences) and 100 μL aliquots were transferred to 5 mL flow cytometry tubes. PE-Annexin V reagent was added to each tube and cells were incubated for 15 minutes at room temperature in dark. Cells were resuspended in 1X binding buffer and PE staining was assessed by flow cytometry. Results were analysed using the FlowJo software (FlowJo LLC. Ashland, OR, USA).

Analysis of the synergistic effect of combination treatment with perifosine and doxorubicin.
T-ALL cells were treated with a solvent control or with increasing concentrations of doxorubicin alone, or in combination with a suboptimal concentration of perifosine (Invivogen) for 24 hours. Cell survival was monitored by resazurin staining. To assess synergistic effects, the CompuSyn software was used to generate combination indices for each drug combination, as previously described 37 . Indices below 0.9 indicate synergism 37 .
Scientific RepoRts | 7: 14767 | DOI:10.1038/s41598-017-15200-3 Xenograft models. Animal protocols were approved by the Animal Research Ethics Board (AREB) at University of Saskatchewan. All in vivo experiments were carried out in accordance with relevant guidelines and regulations. Breeder pairs of NOD-SCID mice were from the Jackson Laboratory. ΔB6-Jurkat and pc3-Jurkat were mixed with Corning Matrigel Matrix (Corning, NY, USA) (3 × 10 6 cells resuspended in 50 μL of PBS were mixed with 50 μL of the matrigel) and injected subcutaneously into the right flanks of experimental animals. Experimental groups injected with pc3-Jurkat or ΔB6-Jurkat cells were of an equal size within each independent experiment (n = 7 or n = 6) and contained equal mixtures of male and female animals. Once tumours reached a measurable size, mice were treated weekly via intravenous injections of doxorubicin at 2 mg/kg or a matching saline control. Tumours were measured using a digital caliper twice a week and tumour volumes were calculated using the following equation: A/2*B 2 , where A was long and B was short diameters of a tumour. All experiments were terminated in accordance with AREB guidelines. Statistical analysis. Until otherwise indicated, student's t-test was performed for statistical analyses. Data are presented as the mean ± standard deviation (SD). Data availability. Original data files are available upon a reasonable request.