Inhibition of calcineurin combined with dasatinib has direct and indirect anti-leukemia effects against BCR-ABL1⁺ leukemia

Abstract

Treatment of BCR-ABL1⁺ leukemia has been revolutionized with the development of tyrosine kinase inhibitors. However, patients with BCR-ABL1⁺ acute lymphoblastic leukemia and subsets of patients with chronic myeloid leukemia are at high risk of relapse despite kinase inhibition therapy, necessitating novel treatment strategies. We previously reported synthetic lethality in BCR-ABL1⁺ leukemia cells by blocking both calcineurin/NFAT signaling and BCR-ABL1, independent of drug efflux inhibition by cyclosporine. Here, using RNA-interference we confirm that calcineurin inhibition sensitizes BCR-ABL1⁺ cells to tyrosine kinase inhibition in vitro. However, when we performed pharmacokinetic and pharmacodynamic studies of dasatinib and cyclosporine in mice, we found that co-administration of cyclosporine increases peak concentrations and the area under the curve of dasatinib, which contributes to the enhanced disease control. We also report the clinical experience of two subjects in whom we observed more hematopoietic toxicity than expected while enrolled in a Phase Ib trial designed to assess the safety and tolerability of adding cyclosporine to dasatinib in humans. Thus, the anti-leukemia benefit of co-administration of cyclosporine and dasatinib is mechanistically pleiotropic, but may not be tolerable, at least as administered in this trial. These data highlight some of the challenges associated with combining targeted agents to treat leukemia.

Introduction

Chronic myeloid leukemia (CML) and some cases of acute lymphoblastic leukemia (Ph⁺ ALL) are caused by the translocation of chromosomes 9 and 22 and the resulting oncogenic fusion gene BCR-ABL1. The fusion of the BCR gene to the ABL gene results in constitutive activation of the ABL kinase with numerous downstream effects (1, 2). Targeting of BCR-ABL1 with tyrosine kinase inhibitors (TKI) has been enormously successful for patients with chronic phase CML. Event free survival at 5 years in patients with chronic phase (CP) CML treated with imatinib mesylate (IM) exceeds 80% (3), significantly higher than that of historical controls treated with interferon and cytarabine (4). Nonetheless, IM therapy fails for a substantial proportion of CP CML patients due to unsatisfactory therapeutic effect (15%) or poor tolerance of the medication (8%) (5). Further, discontinuation of IM results in molecular relapse in more than 50% of patients (6), possibly due to either the relative insensitivity of primitive leukemia progenitor cells to IM (7) or the non-addiction of hematopoietic stem cells to BCR-ABL1 (8). Despite the success of IM and newer TKI (9–11) in CP CML, advanced phase CML and Ph⁺ ALL are not as responsive to TKI monotherapy. In small, early phase studies, IM alone resulted in short-lived complete hematologic remissions in less than 25% of patients with blast phase CML or relapsed/refractory Ph⁺ ALL (12–14). Combining TKI with corticosteroids or conventional chemotherapeutics leads to complete remission (CR) in more than 90% of patients (15, 16), but optimal regimens remain to be determined and the likelihood of relapse remains high for these patients. Thus, there is an urgent need for novel methods of disease control in patients with more aggressive BCR-ABL1⁺ leukemias.

In an effort to identify novel therapeutic strategies, we used a large-scale shRNA library to screen for targets and signaling pathways, which when inhibited, sensitize BCR-ABL1⁺ cells to IM (17). These and subsequent experiments identified a non-canonical Wnt/Calcineurin/NFAT pathway that is critical for BCR-ABL1⁺ cell survival, but only in the presence of tyrosine kinase inhibition. Calcineurin is a central component of this pathway, and is inhibited by cyclosporine. Cyclosporine has been safely used for decades in patients for immunosuppression after organ transplantation. When used in vitro, cyclosporine potently synergizes with imatinib and dasatinib in killing leukemia cells with both myeloid and lymphoid phenotypes. More importantly, in an aggressive murine model of BCR-ABL1⁺ leukemia, combining cyclosporine with dasatinib significantly prolongs leukemia-free survival as compared to dasatinib alone. In fact, mice treated with the combination were apparently cured of their leukemia, while all of the mice treated with dasatinib alone died of leukemia. Studies to elucidate the mechanism of synergy implicate NFAT, the downstream target of calcineurin, as an essential mediator of TKI sensitivity. Importantly, several lines of investigation indicate that the increased sensitivity to IM in the presence of cyclosporine is not due to cyclosporine mediated inhibition of drug efflux pumps (17).

In humans, both cyclosporine and dasatinib are metabolized by the cytochrome p450 enzyme Cyp3A4 (18) and are substrates of p-glycoprotein, raising the possibility of altered pharmacology and metabolism when used in combination. P-glycoprotein and CYP3A functionally interact in three ways (19). First, drugs are repeatedly taken up and pumped out of the enterocytes by P-glycoprotein. Repeated exposure to CYP3A enzymes increases probability of drugs being metabolised. Second, P-glycoprotein keeps intracellular drug concentrations within the linear range of the metabolising capacity of CYP3A enzymes. Third, P-glycoprotein transports drug metabolites formed in the mucosa back into the gut lumen (20, 21).

We sought to determine whether cyclosporine influences dasatinib exposure in mice treated with combination therapy and if altered pharmacokinetics play a role in enhanced survival of combination treated mice. Mice with and without BCR-ABL1⁺ leukemia were treated with varying doses of single agent or combined therapy, and dasatinib and cyclosporine levels were determined by tandem liquid chromatography and mass spectroscopy methods developed in our lab. We demonstrate that cyclosporine does indeed increase dasatinib exposure in mice treated with combination therapy. In parallel, we enrolled patients in a Phase Ib trial to assess the safety and tolerability of adding cyclosporine to dasatinib in humans, and observed more hematopoietic toxicity than anticipated. These data highlight the importance of thorough pre-clinical evaluation in vivo and the challenges associated with combining targeted agents.

Materials and Methods

Cell culture and mouse model of leukemia

Luciferase- or green fluorescence protein (GFP) -expressing, BCR-ABL⁺ Arf^−/− cells were kindly provided by Dr. Richard Williams (22, 23). Cells were cultured in RPMI supplemented with 20% FBS, L-glutamine, β-ME, and antibiotics. Lentiviruses expressing shRNAs and packaging vectors were transfected into 293FT cells to generate viral particles, as previously described (24). Leukemia cells were cultured in virus-containing media for 48 hours and then selected in puromycin. Single cell clones were generated by limiting dilution in 96-well plates. Cells were counted by flow cytometry on a Guava Easy Cyte Plus (Millipore, Billerica, MA), with propidium iodide (Sigma-Aldrich, St. Louis, MO) exclusion. To generate a highly consistent and aggressive leukemia, 5×10⁵ cells were transferred into un-irradiated C57Bl/6 recipients (17). Leukemia burden was tracked over time via IVIS imaging for luciferase expression or by flow cytometry for GFP and lineage markers of bone marrow as previously described (25, 26). Dasatinib (Sprycel; Bristol Myers Squib) and Cyclosporine (Neoral; Novartis) were purchased from the pharmacies of University of Colorado Hospital and Children's Hospital Colorado (both in Aurora, CO). Dasatinib was dissolved in 80mM citric acid (pH 2.1) every 2–3 days. One or both drugs were diluted in vehicle and the medications were administered by oral gavage once daily. Dasatinib doses are indicated in figure legends. Cyclosporine dosing was 25mg/kg/day by gavage. Intraperitoneal cyclosporine was 10mg/kg/day. All animal studies were approved by the Institutional Animal Care and Use Committee.

Antibodies, Reagents and flow cytometry

Information on antibodies for flow cytometry and western blotting is provided in supplemental table 1. shRNA constructs were from the TRC2 collection and obtained from the Functional Genomics Core of the University of Colorado Cancer Center. Specific constructs are listed in Supplemental Table 2. Dasatinib for cell culture experiments was purchased from LC Laboratories (Woburn, MA) and diluted in DMSO (Sigma Aldrich, St. Louis, MO). Other reagents were purchased from Sigma Aldrich. Peripheral blood and bone marrow cells were prepared for flow cytometry as previously described (24). Stained samples were analyzed on either a Guava Easy Cyte Plus (Millipore, Billerica, MA), or Cyan ADP Analyzer (Dako North America, Carpinteria, CA) flow cytometer.

Pharmacokinetic analyses

Peripheral blood samples were frozen at −80°C. All samples from each experiment were thawed and analyzed for dasatinib concentration at the same time. Once thawed, dasatinib concentrations were determined using an assay developed to monitor a structurally similar compound IM by LC/LC-MS/MS (27), with minor modifications. Using this assay, concentrations of dasatinib in whole blood can be determined with a lower limit of detection of 50pg/ml. PK parameters such as half-life (T_½), peak concentration (C_max), and area under the curve (AUC_inf) were calculated using WinNonlin software (Pharsight, St. Louis, MO).

Clinical trial

The Phase Ib trial was supported by the U.S. National Cancer Institute and The Leukemia & Lymphoma Society, registered with clinicaltrials.gov (NCT01426334), approved by the Colorado Multiple Institutional Review Board, and conducted in accordance with the Helsinki Declaration of 1975 as revised in 2008. Dasatinib was provided through an agreement with the Cancer Therapy Evaluation program of the NCI. In brief, patients over 18 years old with histologically or cytologically confirmed CML and without organ dysfunction were eligible for enrollment if they had 1) chronic phase CML and were refractory or intolerant of treatment, or 2) accelerated phase CML for which allogeneic transplantation was not expected to occur within one month of enrollment. After informed consent, subjects were treated with dasatinib for 7 days in order to achieve steady state levels, at which point pharmacokinetic samples were obtained and cyclosporine treatment was added. Cyclosporine trough levels were monitored with goal levels of 125–200ng/ml, a range in which calcineurin is expected to be inhibited, but with minimal risk of toxicity (28). Seven days after this level was achieved, pharmacokinetic samples were again obtained. The duration of therapy was scheduled to be up to 4 months in the absence of specified clinical parameters, withdrawal of consent or the occurrence of specified adverse effects including more than 1 occurrence of Grade 3 hematologic toxicity.

Statistical and flow cytometric analyses

Excel (Microsoft, Redmond, WA) and GraphPad Prism 5 (GraphPad Software, San Diego, CA) were used for data sorting, analysis and graphical depiction of data. Student's t-test was used to compare 2 samples; analysis of variance (ANOVA) and Bonferroni's post-test were used to compare more than 2 samples. The Mantel-Cox (log-rank) test was used to test for significant differences in survival. CalcuSyn was used to determine the combination index (29). Except as indicated in figure legends, figures demonstrate means of data collected from at least 3 independent experiments completed in duplicate or triplicate for each condition. Error bars in each figure depict the standard error of the mean, and may be obscured when narrow.

Results

Calcineurin inhibition sensitizes BCR-ABL1⁺ leukemia cells to dasatinib

We previously demonstrated that the combination of cyclosporine and TKI was synergistic in killing human BCR-ABL1⁺ leukemia cell lines. Importantly, this synergistic effect was independent of drug efflux through inhibition of MDR1 (17). We have also observed synergistic inhibition of engineered murine leukemia cells expressing BCR-ABL1 when treated with dasatinib and cyclosporine (Figure 1A). To confirm that this effect was not an off-target effect of cyclosporine, we sought to inhibit calcineurin genetically. Calcineurin is a phosphatase complex composed of a catalytic subunit (calcineurin A) encoded by 3 genes, and a regulatory subunit (calcineurin B), encoded by only 1 gene (Ppp3r1) in non-germline tissue (30). Thus, knockdown of Ppp3r1 should impair calcineurin function. Indeed, shRNA directed against Ppp3r1 knocked down the regulatory subunit (CNB1) and sensitized these cells to dasatinib (Figure 1B). While the shift in IC₅₀ in this short-term assay was somewhat modest (from 0.81 to 0.46 nM), the effect was amplified in clonal populations exposed to dasatinib for 72 hours and then allowed to recover in the absence of treatment (Figure 1C). Taken together, these data indicate that inhibition of calcineurin, either pharmacologic or genetic, sensitizes BCR-ABL1⁺ cells to dasatinib.

Figure 1. Inhibition of calcineurin sensitizes murine BCR-ABL1⁺ leukemia cells to dasatinib.

A. Murine Arf^−/−/BCR-ABL1⁺ cells were treated with dasatinib and/or cyclosporine at the indicated concentrations for 72 hours, stained with propidium iodide and counted by flow cytometry. Live cell numbers are depicted (left). Combination index values are indicated above respective data points, where values less than 1 indicate synergy. Similarly, data points below the diagonal on the normalized isobologram (right) indicate synergistic inhibition of proliferation. B. Arf^−/−/BCR-ABL1⁺ cells were transduced with non-silencing control shRNA (shNS) or shRNAs directed against the regulatory subunit B of calcineurin (Ppp3r1; shCN-151, −166, −351) and selected in puromycin. Whole cell lysates were subject to western blot for the protein (CN-B1) and tubulin. A representative blot is shown. The graph demonstrates quantification using ImageJ software of CN-B1 expression relative to shNS from 2 different blots. shNS and shCN-351 cells were treated with dasatinib at increasing concentrations and counted 72 hours later. Dose response curves were generated by non-linear regression. C. Knockdown was confirmed by western blot (inset) in clonal populations derived from shNS and shCN-351 pools. These were treated for 72 hours with dasatinib (1nM) or vehicle (DMSO), counted, split 1:10 into media without dasatinib, and counted again after another 72 hours.

Co-administration of cyclosporine with dasatinib increases dasatinib concentration and tyrosine kinase inhibition

While in vitro data indicate cell-autonomous sensitization to tyrosine kinase inhibition, the shared mechanisms of metabolism of cyclosporine and dasatinib raise the possibility of non-cell autonomous mechanisms in vivo, such as altered pharmacokinetics and pharmacodynamics. To determine the effect of combined therapy on the pharmacokinetics of dasatinib, mice without leukemia were treated with cyclosporine, dasatinib or combination therapy at doses previously shown to have differential effects on survival in leukemia-bearing mice (17). Peripheral blood was collected at the indicated time-points and cyclosporine and dasatinib levels determined. At the doses tested, combination therapy resulted in higher C_max and AUC_inf of dasatinib after a single dose (Figure 2A and Table I) and after one week of therapy, when steady state levels are expected (Figure 2B and Table I). Pharmacokinetic (PK) analyses indicate that after 1 week of therapy, co-administration of cyclosporine with dasatinib increases the C_max and AUC_inf of dasatinib as compared to dasatinib alone by 2.6- and 1.9-fold, respectively, at these doses, whereas the effect of dasatinib on cyclosporine concentrations is much more modest (Table I). The PK profiles suggest that co-administration of cyclosporine with dasatinib enhances enteric absorption of dasatinib resulting in a higher C_max, but has little effect on its systemic elimination. To confirm that the enhanced exposure to dasatinib levels resulted in greater tyrosine kinase inhibition in hematopoietic cells in vivo, we measured phosphorylated-Src (P-Src) from the peripheral blood of mice without leukemia, as it is a target of dasatinib and has been validated as a biomarker of dasatinib activity (31). Consistent with the altered pharmacokinetics, the addition of cyclosporine enhanced the inhibition of Src phosphorylation by dasatinib, although maximal P-Src inhibition appeared to be reached with the highest dose of dasatinib alone (Figure 2C). For subsequent experiments, C_max was used as a surrogate marker for total dasatinib exposure and activity.

Figure 2. Co-Administration of cyclosporine and dasatinib enhances the peak concentration and activity of dasatinib.

Mice without leukemia were treated with dasatinib, cyclosporine or both by oral gavage at the indicated doses (n=6/group). A. Serum from peripheral blood was obtained at the indicated time-points after single doses and B. after one week of therapy (steady state). Dasatinib and cyclosporine levels were determined by LC/LC-MS/MS. C. Peripheral blood mononuclear cells from mice without leukemia were isolated 4 hours after treatment with dasatinib at the indicated doses with or without cyclosporine (25 mg/kg). Whole cell lysates were subjected to western blot with antibodies directed against phosphorylated-Src, Src and actin. The extent of Src phosphorylation (ImageJ) relative to untreated mice is depicted graphically. (The reduction in Src protein expression shown was not consistent across experiments.)

Table I.

Pharmacokinetics of dasatinib and cyclosporine after one week of treatment

Treatment	T1/2 (hr)	Cmax(ng/ml)	AUCinf (ng/ml*hr)
Dasatinib PK
Dasatinib	5.0	107.7	487.4
Dasatinib + Cyclosporine	4.1	277.4	916.8
Cyclosporine PK
Cyclosporine	5.9	3,279.5	24,990
Dasatinib + Cyclosporine	5.2	3,378.7	32,181

Altered pharmacokinetics in mice with leukemia influences disease progression and survival

Although these formal PK studies demonstrated enhanced dasatinib exposure when combined with cyclosporine, data compiled from subsequent experiments in mice with leukemia highlight the variability of PK among mice and across experiments, and indicate that this phenomenon is most significant at lower doses of dasatinib (Figure 3A). For example, when dasatinib is dosed at 10 mg/kg/d, cyclosporine enhances the mean C_max of dasatinib by 7-fold (93.2 v. 650.0 ng/ml), compared with a non-significant 1.2-fold increase (408.6 v. 495.4 ng/ml) when given at 20 mg/kg/d. This is consistent with the pharmacodynamic data presented in Figure 2C and suggests that near maximal absorption of dasatinib occurs when dasatinib is given at 20 mg/kg/d alone, and at 10 mg/kg/d when given in combination with cyclosporine. To determine whether disease burden correlated with peak concentrations of dasatinib, we measured the percentage of GFP⁺ leukemia cells in the bone marrow of mice after one week of dasatinib or combination therapy. Not surprisingly, residual disease burden in individual mice correlated with the peak concentration of dasatinib (Figure 3B). Similar results were observed in recipients of luciferase expressing leukemia monitored non-invasively (not shown). In long-term experiments, the greater dasatinib exposure associated with co-administration of cyclosporine significantly prolonged survival as compared to dasatinib alone when given at the same dose (Figure 3C; dasatinib 10 mg/kg/d v. cyclosporine plus dasatinib 10 mg/kg/d). At near equivalent dasatinib exposures (dasatinib 10 mg/kg/d v. dasatinib 2 mg/kg/d plus cyclosporine), the median survival in the combination-treated mice was greater (64 v. 48 days), although this difference was not statistically significant (Figure 3C; P=0.3). In order to bypass the local effects of cyclosporine on intestinal absorption, we next delivered the cyclosporine by intraperitoneal injection (IP). Surprisingly, higher concentrations of dasatinib were observed whether cyclosporine was given by oral gavage or IP, which again correlated with prolonged survival (Figure 3D).

Figure 3. Higher dasatinib concentrations due to co-administration of cyclosporine are associated with enhanced leukemia control.

Unirradiated mice were inoculated with 5×10⁵ GFP⁺ or luciferase⁺ Arf^−/−/BCR-ABL⁺ leukemia cells. Mice were subsequently treated with dasatinib at the indicated doses (mg/kg/d) with or without cyclosporine (25mg/kg/d). A. After 7 days of therapy by oral gavage, peripheral blood was obtained 45 minutes after drug dosing (C_max) and analyzed by LC/LC-MS/MS for dasatinib concentrations (Data are compiled from 6 independent experiments; n≥10 for each group). (Asterisks indicate P<0.05; ANOVA with Bonferroni's post-test. NS=non-significant) B. After 7 days of therapy by oral gavage the BM was collected for analyses by flow cytometry. The BM leukemia burden (GFP⁺/B220⁺ cells) and serum dasatinib concentrations of each mouse from 2 independent experiments are depicted. R² and P values are from Pearson's correlation test. Representative dot-plots from flow cytometry are depicted. C. Dasatinib C_max from mice with leukemia treated by oral gavage in an individual experiment are demonstrated. Mice were euthanized when ill appearing and overall survival is depicted on the Kaplan-Meier curve. D. Dasatinib C_max from mice with leukemia treated with dasatinib by oral gavage and cyclosporine by intraperitoneal injection (10mg/kg/d) in an individual experiment are shown. Mice were euthanized when ill appearing and overall survival is depicted on the Kaplan-Meier curve.

Cyclosporine may enhance pharmacokinetics and toxicity of dasatinib in humans

In parallel with our laboratory studies, we opened a Phase Ib trial (NCT01426334), to determine the safety and tolerability of combining dasatinib and cyclosporine for select patients with BCR-ABL1⁺leukemia. After informed consent, subjects were treated with dasatinib for 1 week, after which they were treated with dasatinib and cyclosporine. Here we present the clinical experience of 2 of the subjects. From each, peripheral blood dasatinib concentrations were measured at specified time-points, both when on dasatinib alone and when given in combination with cyclosporine.

UPIN001 was a 48 year-old female with accelerated phase chronic myeloid leukemia (CML), who had been previously treated with imatinib and nilotinib without hematologic remission. Her leukemia cells expressed BCR-ABL1 with Y253H mutation (32). She was enrolled on the trial after being started on dasatinib 140mg daily. Cyclosporine 200 mg twice daily was initiated after one week on study with adjustments in dosage to maintain troughs between 125–200ng/ml. For UPIN001, the addition of cyclosporine appeared to increase the C_max of dasatinib, as compared to when she was on dasatinib alone (Figure 4A). After about 4 weeks of combination therapy, she had Grade 1 thrombocytopenia and Grade 3 neutropenia (Figure 4B, C), necessitating temporary interruption of her dasatinib. After being restarted on dasatinib at a reduced dose of 100mg daily, she again had Grade 3 neutropenia and Grade 4 thrombocytopenia, prompting her removal from study due to protocol-defined toxicity parameters. UPIN002 was a 29 year-old male with newly diagnosed chronic phase CML, who was treated only with hydroxyurea for hyperleukocytosis prior to being enrolled in the trial. His leukemia cells expressed BCR-ABL1 with a small insertion of unknown significance (33). He was started on dasatinib 100mg daily, and cyclosporine 125 mg twice daily was initiated one week later. The peak concentration of dasatinib was only slightly increased in UPIN002 after the addition of cyclosporine (Figure 4D), but he also developed Grade 2 thrombocytopenia (Figure 4E) that persisted for several weeks, leading to his eventual elective withdrawal from the trial. While these data are not conclusive, we suspect that the co-administration of cyclosporine enhances the absorption of dasatinib and increases its hematologic toxicity.

Figure 4. Co-administration of cyclosporine with dasatinib may enhance serum concentrations and toxicity of dasatinib in patients.

Peripheral blood from 2 patients enrolled in a Phase Ib trial was analyzed by LC/LC-MS/MS for dasatinib concentrations at the indicated time points after 7 days of therapy with dasatinib only and again after at least 7 days of dasatinib and cyclosporine at a dose that resulted in serum cyclosporine concentrations between 125–200ng/ml. Complete blood counts were monitored periodically for the duration of subjects' participation in the trial. A, B & C. Dasatinib concentrations, platelet (Plt) and absolute neutrophil counts (ANC) from UPIN001 are depicted. D, E, F. Dasatinib concentrations, Plt counts and ANC from UPIN002.

Discussion

While treatment of BCR-ABL1⁺ leukemias has been revolutionized with the development of TKI, there remain populations for whom this strategy is insufficient and novel strategies are necessary. We recently identified a non-canonical Wnt/Calcineurin/NFAT signaling cascade that serves as an escape pathway mechanism in BCR-ABL1⁺ cells upon exposure to tyrosine kinase inhibition (17). While we demonstrated that the synergistic effect of inhibiting calcineurin with cyclosporine was not dependent upon enhanced tyrosine kinase inhibition in vitro, we had not examined the pharmacokinetic or pharmacodynamic effects of combinatorial therapy in vivo. Here, we report that co-administration of cyclosporine with dasatinib increases the peak plasma concentrations of dasatinib, and that the increased plasma concentrations are associated with greater tyrosine kinase inhibition, as well as disease control. Taken with the demonstration that knockdown of calcineurin sensitizes BCR-ABL1⁺ cells to dasatinib in vitro along with our previously published data, we conclude that the survival benefit in mice with leukemia treated with cyclosporine and dasatinib is attributable to at least two mechanisms: enhanced tyrosine kinase inhibition due to higher plasma TKI concentrations and inhibition of an escape pathway. In 2 subjects treated in the clinical trial performed in parallel with these laboratory studies, we observed more hematopoietic toxicity than we expected, which we speculate may be due to higher plasma concentrations of dasatinib.

In our previous work, we demonstrated enhanced human BCR-ABL1⁺ cell killing by imatinib and dasatinib by inhibiting FZD-8 and CAMKII genetically, and CAMKII and calcineurin pharmacologically, implicating the non-canonical Wnt/calcineurin/NFAT signaling cascade as a process through which BCR-ABL1⁺ cells escape tyrosine kinase inhibition. Importantly, we did not detect enhanced reduction in tyrosine phosphorylation in the K562 cells treated with cyclosporine, excluding enhancement of tyrosine kinase inhibition, such as by reduced drug efflux. Here, we knocked down calcineurin and saw similar sensitization to TKI in a murine BCR-ABL⁺ leukemia, confirming the importance of this pathway in mediating sensitivity to TKI.

Nonetheless, we hypothesized that because of overlapping processing pathways, including p-glycoprotein and CYP3A4 (in humans), co-administration of cyclosporine with dasatinib may influence the pharmacokinetics of either or both of the drugs in vivo. We found that the peak concentration of dasatinib was increased with co-administration of cyclosporine, likely due to competitive inhibition of drug efflux and metabolism processes. Whether similar effects would be observed with tacrolimus, another FDA approved calcineurin inhibitor, remains to be determined, but might be predicted as it is also inhibits p-glycoprotein (34).

As we had observed enhancement of survival in mice with leukemia due to treatment with two FDA approved drugs (17), we designed a Phase Ib clinical trial to test the safety and tolerability of combining these agents in select patients with BCR-ABL1⁺ CML. We were surprised to observe precipitous drops in the platelet counts and absolute neutrophil counts of the two patients treated with both cyclosporine and dasatinib. This toxicity seemed greater than expected with dasatinib alone, and in light of the pharmacokinetics in mice, we presume is a consequence of the co-administration of cyclosporine. Due the possibility of enhanced toxicity and expected slow accrual because of emerging data on second generation TKI (9, 10), we elected to close the trial.

These data highlight the importance of thorough in vivo modeling of combination therapeutic strategies identified in the laboratory as they are translated into the clinic. While our data strongly implicate calcineurin as a key mediator of BCR-ABL1⁺ leukemia cell survival in the face of tyrosine kinase inhibition, we do not advocate for combination therapy outside of the context of a clinical trial because of potential toxicities. Alternative dosing regimens of cyclosporine or other calcineurin inhibitors will need to be studied in a multi-center trial in order to draw meaningful conclusions about this therapeutic strategy.