Inhibition of Xanthine Oxidoreductase Enhances the Potential of Tyrosine Kinase Inhibitors against Chronic Myeloid Leukemia

Chronic myeloid leukemia (CML) is characterized by the expression of the oncogenic kinase BCR-ABL. Although tyrosine kinase inhibitors (TKIs) against BCR-ABL represent the standard therapeutic option for CML, resistances to TKIs can be a serious problem. Thus, the search for novel therapeutic approaches is still needed. CML cells show an increased ROS production, which is required for maintaining the BCR-ABL signaling cascade active. In line with that, reducing ROS levels could be an interesting therapeutic strategy for the clinical management of resistant CML. To analyze the therapeutic potential of xanthine oxidoreductase (XOR) in CML, we tested the effect of XOR inhibitor allopurinol. Here, we show for the first time the therapeutic potential of allopurinol against BCR-ABL-positive CML cells. Allopurinol reduces the proliferation and clonogenic ability of the CML model cell lines K562 and KCL22. More importantly, the combination of allopurinol with imatinib or nilotinib reduced cell proliferation in a synergistic manner. Moreover, the co-treatment arms hampered cell clonogenic capacity and induced cell death more strongly than each single-agent arm. The reduction of intracellular ROS levels and the attenuation of the BCR-ABL signaling cascade may explain these effects. Finally, the self-renewal potential of primary bone marrow cells from CML patients was also severely reduced especially by the combination of allopurinol with TKIs. In summary, here we show that XOR inhibition is an interesting therapeutic option for CML, which can enhance the effectiveness of the TKIs currently used in clinics.


Introduction
Chronic myeloid leukemia (CML) is a hematological malignancy originated from the chromosomal translocation t(9,22)(q34;q11) that produces the Philadelphia chromosome [1]. As a result of this translocation, the oncogenic kinase BCR-ABL is expressed. This constitutively active kinase is capable of turning on several signaling pathways, including PI3K/AKT, STAT5, MAPKs, allowing growth factor-independent cell proliferation and escape of apoptosis [2]. With the discovery of specific tyrosine kinase inhibitors (TKIs) against BCR-ABL, the therapy of CML changed dramatically from a dismal to a very favorable outcome [3]. However, primary or secondary resistance to these treatments is still a serious threat for CML patients [4], which justifies the search for novel therapeutic options. Fraction affected or percentage of inhibition with respect to control in KCL22 cells (n = 6). * p < 0.001 and *** p < 0.05 reflect significant differences with respect to control.

Allopurinol and TKIs Inhibits K562 and KCL22 Cells Proliferation in a Synergistic Manner
Bearing in mind that CML cells are sensitive to allopurinol treatment (Figure 1), we next combined imatinib and allopurinol ( Figure 2). The inhibition of cell proliferation was significantly more pronounced with the combination in both K562 ( Figure 2a) and KCL22 cells (Figure 2b). Moreover, the analysis of drug interaction showed CIs significantly below 1, thus reflecting a strong synergistic effect of the allopurinol plus imatinib combinations in both cell lines (Figure 2c,d). (b) Fraction affected or percentage of inhibition with respect to control in KCL22 cells (n = 6). * p < 0.001 and *** p < 0.05 reflect significant differences with respect to control.

Allopurinol and TKIs Inhibits K562 and KCL22 Cells Proliferation in a Synergistic Manner
Bearing in mind that CML cells are sensitive to allopurinol treatment (Figure 1), we next combined imatinib and allopurinol ( Figure 2). The inhibition of cell proliferation was significantly more pronounced with the combination in both K562 ( Figure 2a) and KCL22 cells (Figure 2b). Moreover, the analysis of drug interaction showed CIs significantly below 1, thus reflecting a strong synergistic effect of the allopurinol plus imatinib combinations in both cell lines (Figure 2c,d).  To test this further, the combination of a second-generation BCR-ABL inhibitor, nilotinib, with allopurinol was also tested. In agreement with the results described above, the combination reduced the proliferation more strongly than the individual treatments (Figure 3a,b). Moreover, the analysis of the drug interaction supports the synergism between both drugs, as the CI was below 1 (Figure  3c,d). The combination of imatinib and allopurinol reduces the proliferation of the K562 and KCL22 cells in a synergistic manner. K562 and KCL22 cells were treated with different concentrations of imatinib, allopurinol or their combination for 48 h. Proliferation was analyzed by MTT assay and the combination indexes (CI) were calculated as described in the Methods section. (a) K562 cells proliferation with respect to control (n = 4). (b) KCL22 cells proliferation with respect to control (n = 5). Significant differences: *** p < 0.001, ** p < 0.01 with respect to control; +++ p < 0.001, ++ p < 0.01 with respect to allopurinol-treated cells; ### p < 0.001, # p < 0.05 with respect to imatinib-treated cells. (c) Mean CI values for the drug combinations tested in K562 cells (n = 4). (d) Mean CI values for the drug combinations tested in KCL22 cells (n = 5). *** p < 0.001 and ** p < 0.01 reflect significant differences with respect to CI value 1.
To test this further, the combination of a second-generation BCR-ABL inhibitor, nilotinib, with allopurinol was also tested. In agreement with the results described above, the combination reduced the proliferation more strongly than the individual treatments (Figure 3a,b). Moreover, the analysis of the drug interaction supports the synergism between both drugs, as the CI was below 1 (Figure 3c,d).
In line with the results described in Figures 2 and 3, a stronger reduction of viable cell numbers by the co-treatment arms with respect to single arms was also observed by cell counting with trypan blue exclusion ( Figure 4).  In line with the results described in Figures 2 and 3, a stronger reduction of viable cell numbers by the co-treatment arms with respect to single arms was also observed by cell counting with trypan blue exclusion ( Figure 4).    In line with the results described in Figures 2 and 3, a stronger reduction of viable cell numbers by the co-treatment arms with respect to single arms was also observed by cell counting with trypan blue exclusion ( Figure 4).  cells with respect to control (n = 5). Significant differences: *** p < 0.001, ** p < 0.01 with respect to control; ++ p < 0.01 with respect to allopurinol-treated cells; ## p < 0.01, # p < 0.05 with respect to TKI-treated cells.
We next analyzed the effect on cell viability. While the single-treatment with each drug induced a subtle decrease in the percentages of viable cells, the most pronounced effect was observed upon the co-treatment arms (allopurinol + TKI) in both model cell lines ( Figure 5 and Tables 1 and 2). The combination induced a stronger decrease in the number of viable cells and in an increase of apoptotic cells ( Figure 5, Tables 1 and 2). Therefore, the addition of allopurinol to either imatinib or nilotinib induces cell death more efficiently than individual treatments.

Allopurinol and TKIs Combination Reduces K562 and KCL22 Cells Clonogenic Capacity
The effect of anti-leukemic drugs on cell renewal capacity is an important aspect to analyze. This can be done through colony-forming unit assays. Allopurinol treatment reduced the clonogenic capacity of both K562 and KCL22 cells (Figure 6), thereby supporting again the potential use of allopurinol against CML. At the concentration used, both TKIs (imatinib and nilotinib) reduced the clonogenic ability of K562 cells (Figure 6a,c), despite the fact that no significant effect was seen in KCL22 cells. However, the combinations of allopurinol + TKI showed the most potent effect. Allopurinol + nilotinib co-treatment was the most effective treatment, leading to a significant decrease in colonies when comparing either to the control or to the single treatment (Figure 6c,d). This evidence, in line with the results described above, supports the benefit of adding allopurinol to TKIs. Significant differences: *** p < 0.001, ** p < 0.01, * p < 0.05 with respect to control; + p < 0.05 with respect to allopurinol-treated cells; ## p < 0.01, # p < 0.05 with respect to TKIs-treated cells.

Imatinib, Allopurinol, and Their Combination Reduce Intracellular ROS Levels
ROS are important for CML progression, as they facilitate BCR-ABL signaling [19] and increase genetic instability which can eventually lead to progression into the blastic phase of the disease [17]. We reasoned that analyzing the intracellular ROS levels upon the different treatments could help to explain the effects described in the previous sections. In agreement with previous reports [19,33], imatinib treatment reduced the level of intracellular ROS (Figure 7). XOR inhibition with allopurinol also induced a significant reduction in the level of intracellular ROS. The combination of imatinib/allopurinol caused the strongest reduction with respect to the control (Figure 7). While individual treatments induce a decrease of 20%, the combination treatment almost reaches a 40% reduction, suggesting an additive effect of both agents regarding the reduction in ROS levels.  4). (c) CFU number with respect to control in K562 treated with nilotinib, allopurinol or their combination (n = 5). (d) CFU number with respect to control in KCL22 treated with nilotinib, allopurinol or their combination (n = 4). Significant differences: *** p < 0.001, ** p < 0.01, * p < 0.05 with respect to control; + p < 0.05 with respect to allopurinol-treated cells; ## p < 0.01, # p < 0.05 with respect to TKIs-treated cells.

Imatinib, Allopurinol, and Their Combination Reduce Intracellular ROS Levels
ROS are important for CML progression, as they facilitate BCR-ABL signaling [19] and increase genetic instability which can eventually lead to progression into the blastic phase of the disease [17]. We reasoned that analyzing the intracellular ROS levels upon the different treatments could help to explain the effects described in the previous sections. In agreement with previous reports [19,33], imatinib treatment reduced the level of intracellular ROS (Figure 7). XOR inhibition with allopurinol also induced a significant reduction in the level of intracellular ROS. The combination of imatinib/allopurinol caused the strongest reduction with respect to the control (Figure 7). While individual treatments induce a decrease of 20%, the combination treatment almost reaches a 40% reduction, suggesting an additive effect of both agents regarding the reduction in ROS levels.
explain the effects described in the previous sections. In agreement with previous reports [19,33], imatinib treatment reduced the level of intracellular ROS (Figure 7). XOR inhibition with allopurinol also induced a significant reduction in the level of intracellular ROS. The combination of imatinib/allopurinol caused the strongest reduction with respect to the control (Figure 7). While individual treatments induce a decrease of 20%, the combination treatment almost reaches a 40% reduction, suggesting an additive effect of both agents regarding the reduction in ROS levels.

Imatinib, Allopurinol and Their Combination Attenuate BCR-ABL Signaling
The BCR-ABL signaling cascade upon the different treatments was analyzed next. By western blotting, we analyzed the level of activation of BCR-ABL itself, and also the level of activation of STAT5, a direct target of BCR-ABL, and a prominent driver of CML [34]. Imatinib treatment induced a decrease in both the level of activated BCR-ABL (phosphorylated form) and in the total protein levels (Figure 8a), which was in line with our previous results [19]. Interestingly, a very similar result was found upon allopurinol treatment (Figure 8a). The combination of both agents showed a significant decrease in the level of both, the activated form of BCR-ABL and in the total levels of the protein. Such a decrease in BCR-ABL could explain the results we have found: the inhibition of proliferation, an increase of cell death and a reduced clonogenic capacity. DCFDA mean fluorescence intensity with respect to control reflecting intracellular ROS levels is shown (n = 8). *** p < 0.001, * p < 0.05 reflects significant differences with respect to control.

Imatinib, Allopurinol and Their Combination Attenuate BCR-ABL Signaling
The BCR-ABL signaling cascade upon the different treatments was analyzed next. By western blotting, we analyzed the level of activation of BCR-ABL itself, and also the level of activation of STAT5, a direct target of BCR-ABL, and a prominent driver of CML [34]. Imatinib treatment induced a decrease in both the level of activated BCR-ABL (phosphorylated form) and in the total protein levels (Figure 8a), which was in line with our previous results [19]. Interestingly, a very similar result was found upon allopurinol treatment (Figure 8a). The combination of both agents showed a significant decrease in the level of both, the activated form of BCR-ABL and in the total levels of the protein. Such a decrease in BCR-ABL could explain the results we have found: the inhibition of proliferation, an increase of cell death and a reduced clonogenic capacity. When we analyzed STAT5, a characteristic hyper-activated target in CML cells, no differences were found regarding the global levels of this signaling protein. The phosphorylated active form of STAT5 was significantly reduced by imatinib treatment while allopurinol showed the opposite effect. However, the strongest reduction of STAT5 activation was observed upon the co-treatment arm (Figure 8b). When we analyzed STAT5, a characteristic hyper-activated target in CML cells, no differences were found regarding the global levels of this signaling protein. The phosphorylated active form of STAT5 was significantly reduced by imatinib treatment while allopurinol showed the opposite effect. However, the strongest reduction of STAT5 activation was observed upon the co-treatment arm (Figure 8b).

The Combination of Allopurinol and TKIs Reduces the Clonogenic Ability of CML Primary Cells
Finally, to analyze the clinical potential of the combinations tested on cell lines, the clonogenic ability of BM-MNC from CML patients in the presence of the different treatments was studied. Interestingly, the results obtained were very similar to those described above for K562 and KCL22 cells ( Figure 6). Allopurinol or TKIs individually reduced CFU numbers in all patients, but the strongest effect was observed upon the combinations, which led to a significant decrease in colonies when comparing either to the control or to the single drugs ( Figure 9). These observations validate the results described above for CML cell lines, and more importantly, support the feasibility of adding allopurinol to increase the effect of TKIs on CML patients.

Discussion
Eukaryotic cells must cope with the continuous formation of ROS, derived from their aerobic metabolism. ROS have traditionally been considered harmful for cell physiology [35], however, over the last two decades, accumulated evidence supports the importance of a moderate production of ROS for the control of cellular signaling and gene expression [13]. It is accepted that an uncontrolled production of ROS is related to aging and to the development of degenerative diseases and cancer [6,11]. Cancer cells show an elevated level of ROS compared to healthy cells [36,37], a factor that may contribute to tumorigenesis and cancer progression by different mechanisms [38,39]. ROS can damage DNA, increasing the mutation rate [40], thus affecting epigenetic modifications [41], or modifying the activity of different transcription factors involved in cancer [42]. Finally, the transforming activity of several oncogenes, such as KRAS [14,43] or BCR-ABL [16] depends on the upregulated production of ROS.
Increasing ROS levels have been postulated as an attractive therapeutic strategy against cancer [10]. Reducing ROS levels could also be an appealing therapeutic approach, given the importance of ROS in sustaining tumor growth. However, the use of antioxidants in oncology is still a matter of Figure 9. The combination of TKIs and allopurinol reduces the clonogenic capacity of primary chronic myeloid leukemia (CML) cells. BM-MNCs collected from CML patients were treated with TKIs (imatinib and nilotinib), allopurinol, or their combinations for 48 h. The clonogenic capacity after the treatments was analyzed in a semisolid medium by CFU assays. Relative colony numbers of BM-MNCs treated with TKI (imatinib or nilotinib), allopurinol, or their combination are shown. *** p < 0.001, * p < 0.05 with respect to control; ++ p < 0.01, + p < 0.05 with respect allopurinol-treated cells; # p < 0.05 with respect to TKI-treated cells. (n = 5).

Discussion
Eukaryotic cells must cope with the continuous formation of ROS, derived from their aerobic metabolism. ROS have traditionally been considered harmful for cell physiology [35], however, over the last two decades, accumulated evidence supports the importance of a moderate production of ROS for the control of cellular signaling and gene expression [13]. It is accepted that an uncontrolled production of ROS is related to aging and to the development of degenerative diseases and cancer [6,11]. Cancer cells show an elevated level of ROS compared to healthy cells [36,37], a factor that may contribute to tumorigenesis and cancer progression by different mechanisms [38,39]. ROS can damage DNA, increasing the mutation rate [40], thus affecting epigenetic modifications [41], or modifying the activity of different transcription factors involved in cancer [42]. Finally, the transforming activity of several oncogenes, such as KRAS [14,43] or BCR-ABL [16] depends on the upregulated production of ROS.
Increasing ROS levels have been postulated as an attractive therapeutic strategy against cancer [10]. Reducing ROS levels could also be an appealing therapeutic approach, given the importance of ROS in sustaining tumor growth. However, the use of antioxidants in oncology is still a matter of debate that requires further evaluation [44]. Using antioxidants as a co-adjuvant therapy may reduce the toxic side effects produced by pro-oxidants drugs [45], while at the same time antioxidants can reduce the cytotoxicity of many chemotherapeutics.
An alternative to reducing ROS by the use of antioxidants is the inhibition of ROS production sources. This strategy might be more effective and specific if we knew the origin of ROS in tumor cells. NADPH oxidases are one of the main sources of ROS in the cell [12], and they may be an important source of ROS in cancer [46]. We have shown the importance of NADPH oxidase ROS production to maintain BCR-ABL signaling. Moreover, our results in vitro and in vivo show that NADPH oxidases are a potential therapeutic alternative against CML alone or in combination with TKIs [19].
The downregulation and upregulation of XOR activity have been related to different types of solid tumors [47], suggesting the importance of XOR produced ROS in tumorigenesis [48]. Besides, a higher XOR activity has been related with relapsed acute myeloid leukemia [49]. XOR has been used before as a tool to increase oxidative stress in tumor cells as a therapeutic strategy, both in hematological malignancies and solid tumors [48]. Here we have tested the alternative approach, the inhibition of XOR activity against CML. Our results show that allopurinol reduces the proliferation and the clonogenic ability of CML cells, suggesting the feasibility of targeting XOR in CML treatment. More importantly, allopurinol and TKI combination showed a prominent synergistic effect inhibiting CML cell line proliferation. The results obtained in bone marrow cells from CML patients also support the potential of allopurinol to increase the effect of TKIs. While allopurinol and TKIs individually reduced CFU numbers, the combinations showed the strongest effect. Therefore, XOR inhibition could be harnessed to increase the efficacy of the TKIs currently used in the clinic. Although the use of TKIs against CML is one of the best examples of molecular targeted therapy, primary or secondary TKI resistances are still a serious concern [4]. Searching alternatives to enhance TKI effect is a worthwhile endeavor. This would allow the use of lower TKIs doses, which would hamper the appearance of TKIs resistances. In this line, our results offer a very promising strategy by using the XOR inhibitor allopurinol in combination with BCR-ABL inhibitors. Another important aspect of our results is the fact that allopurinol has long been used for the prevention and treatment of gout, which guarantees the safety of human treatment, in addition to its reduced financial cost. Furthermore, many patients at diagnosis combine allopurinol with a TKI to prevent tumor lysis syndrome (TLS), although the former is usually subsequently discontinued [50].
The inhibition of BCR-ABL phosphorylation is the base of the TKIs' mechanism of action. Here we show that allopurinol enhances the potential of TKIs to avoid BCR-ABL activation. This is probably one of the main molecular mechanisms explaining the effectiveness observed upon the combination. In line with our previous results [19], we suggest that a reduction in the intracellular ROS levels by XOR inhibition could hamper BCR-ABL signaling cascade activation. There is an interesting report showing that XOR can enhance NADPH oxidase activity by upregulating cytosolic calcium concentration, which would lead to a further increase in ROS [21]. Therefore, it is tempting to speculate that the effects of allopurinol against CML cells reported here, may well be due to the reduction of ROS production, not only by XOR but also by NADPH oxidases.
The inhibition of STAT5 has been suggested as a feasible therapeutic strategy against CML [51]. Upon co-treatment, we observed a significant decrease in STAT5 activation, which could contribute to the synergistic effect described here. However, allopurinol alone induced a notable increase in the levels of the active form of STAT5. The upregulation of STAT3 in myocardial cells upon allopurinol treatment, probably through JAK2 activation, has been previously described, [52]. In line with this report, we hypothesized that the upregulation of STAT5 activation by allopurinol could also be due to JAK2 activation. The inhibitory effect of the combined therapy on STAT5 activation could thus be explained by the ability of imatinib to inhibit STAT5 activation through JAK2 dependent and independent mechanisms [34]. A similar effect has been reported for the combination of PIM kinases inhibitors with TKIs [53], while PIM kinases inhibitors induced the upregulation of STAT5, the combination of these agents with TKIs led to a strong decrease in STAT5 activation. In addition, during hypoxia the JAK2 /STATs pathway induces XOR activation [54], therefore, it is tempting to speculate that the increase in STAT5 phosphorylation upon allopurinol treatment could be a feedback loop, similar to the one reported for PIM kinases, whose inhibition in combination with TKIs resulted in a synergistic impairment of leukemic cell proliferation [53].
A straightforward consequence of allopurinol treatment is the decrease in intracellular uric acid concentration. Uric acid can react with ROS, NO, and RNS (reactive nitrogen species) triggering oxidation processes that can support cell transformation [55]. Reduction of uric acid level by allopurinol could also contribute to lower intracellular ROS levels, and therefore can be also regarded as an important aspect in the mechanism of action leading to the inhibition of cell proliferation described here. In addition, the existence of ROS independent mechanisms linked to the reduction of cellular uric acid levels cannot be ruled out at this stage, and it will be an interesting aspect to analyze in future studies.
Tumor lysis syndrome (TLS) can be a serious cancer complication derived from excessive cell death, which can eventually lead to organ dysfunction. TLS is characterized by serum hyperuricemia and reducing uric acid levels is regarded as a commonly used prophylactic strategy [56]. Noteworthy, as already mentioned, the use of allopurinol is recommended to minimize TLS associated complications in CML patients [50]. Thus, this strategy could easily be implemented in the clinic and allopurinol could be maintained in the long term, hypothetically increasing the potential achievement of a deep molecular response, one of the main goals in CML treatment.
In the future, it would be interesting to test the combination of XOR inhibitors and TKIs in clinical trials, as well as to test the effect of other recently discovered XOR inhibitors, such as febuxostat, which is more powerful and stable than allopurinol. Its use in the USA has recently been approved by the FDA [57].

Conclusions
Our results show that XOR inhibition with allopurinol reduces CML cell proliferation and clonogenic capacity. Moreover, the allopurinol and TKIs combinations were significantly more effective than the individual drugs regarding the inhibition of cell proliferation or clonogenic capacity as well as in the induction of cell death. Analysis of drug interaction by the median-effect method as described by Chou-Talalay [24] rendered CIs below 1, supporting the synergistic effect of the TKI plus allopurinol combination on inhibiting cell proliferation. We suggest that these effects are due to the reduction of the intracellular ROS content, which leads to the inhibition of the BCR-ABL signaling cascade. In summary, our results offer a simple, safe, and inexpensive potential therapeutic intervention for CML, which could enhance the effectiveness of TKIs, contribute to the achievement of deep molecular responses, and minimize the possibility of resistance and/or progression.