Letters to the Editor 654 used as conditioning regimen before autologous transplantation 2Department of Experimental Immunology, Academic 5 Medical Center, Amsterdam, The Netherlands and in chemo-refractory CLL. In addition, a phase I–II trial in which 3 a comparable regimen (OFAR; consisting of oxaliplatin, Department of Hematology, Maastricht fludarabine, cytarabine and rituximab) was used, yielded University Medical Center, Maastricht, The Netherlands moderate though encouraging results in chemo-refractory E-mail: [email protected] patients; 33% patients with fludarabine-resistant CLL (and 37% of patients with documented deletion of 17p) responded.6 Together these results provide a rationale to explore the References efficacy and mechanism of action of platinum-based regimens in this patient category. A drawback of regimens containing 1 Keating MJ, Flinn I, Jain V, Binet JL, Hillmen P, Byrd J et al. platinum-based compounds is the rather substantial risk of Therapeutic role of alemtuzumab (Campath-1H) in patients who have failed fludarabine: results of a large international study. Blood hematological toxicity, especially in patients with already 2002; 99: 3554–3561. compromised hematopoiesis because of the disease. Therefore, 2 Tam CS, Wierda WG, O’Brien S, Lerner S, Khouri IF, Kantarjian HM the potential benefit of alternative regimens like the combina- et al. Life after fludarabine, cyclophosphamide, & rituximab tion of high-dose methylprednisolone and rituximab, which (FCR)Fthe clinical outcome of patients with chronic lymphocytic showed considerable activity (an overall response rate of 93%) leukemia who receive salvage treatment after frontline FCR. Blood in a small study in a similar patient group, and the newer 2008; 112: 727. 7 3 Vellenga E, van Putten WL, van ‘t Veer MB, Zijlstra JM, Fibbe WE, targeted compounds (reviewed by Tsimberidou and Keating ), van Oers MH et al. Rituximab improves the treatment results of should be explored unabatedly. DHAP-VIM-DHAP and ASCT in relapsed/progressive aggressive CD20+ NHL: a prospective randomized HOVON trial. Blood 2008; 111: 537–543. Conflict of interest 4 Moufarij MA, Sampath D, Keating MJ, Plunkett W. Fludarabine increases oxaliplatin cytotoxicity in normal and chronic lymphocy- The authors declare no conflict of interest. tic leukemia lymphocytes by suppressing interstrand DNA crosslink removal. Blood 2006; 108: 4187–4193. 5 Majolino I, Ladetto M, Locasciulli A, Drandi D, Benedetti F, Acknowledgements Gallamini A et al. High-risk fludarabine-pretreated B-cell chronic lymphocytic Leukemia’s high response rate following sequential We thank GA Huls, MD PhD (University Medical Center, DHAP and alemtuzumab administration though in absence of Groningen), S Hovenga, MD PhD (Nij Smellinghe Hospital, molecular remission. Med Oncol 2006; 23: 359–368. Drachten) and SH Wittebol, MD (Meander Medical Center, 6 Tsimberidou AM, Wierda WG, Plunkett W, Kurzrock R, O’Brien S, Wen S et al. Phase I–II study of oxaliplatin, fludarabine, cytarabine, Amersfoort) for providing patient data. APK is personally and rituximab combination therapy in patients with Richter’s supported by a ‘Veni’-grant from the Netherlands Organization syndrome or fludarabine-refractory chronic lymphocytic leukemia. for Health Research and Development. J Clin Oncol 2008; 26: 196–203. 1,2 3 2 1 7 Tsimberidou AM, Keating MJ. Treatment of fludarabine-refractory SH Tonino , M van Gelder , E Eldering , MH van Oers chronic lymphocytic leukemia. Cancer 2009; 115: 2824–2836. 1 and AP Kater 8 Kaspers GJ, Veerman AJ, Pieters R, Van Zantwijk I, Hahlen K, 1 Department of Hematology, Academic Medical Center, van Wering ER. Drug combination testing in acute lymphoblastic Amsterdam, The Netherlands; leukemia using the MTT assay. Leuk Res 1995; 19: 175–181.

NUP98–HMGB3: a novel oncogenic fusion

Leukemia (2010) 24, 654–658; doi:10.1038/leu.2009.241; ling and possibly of ’s interaction with published online 3 December 2009 DNA (for a review, see Travers6). The expression of the murine Hmgb3 is tightly regulated during hema- NUP98 is a promiscuous gene involved in chromosomal topoiesis and is required in early steps of hematopoietic aberrations with more than 20 different partner in a stem cell development in which it regulates cell-fate decisions. variety of human hematological malignancies.1,2 These rearran- It is expressed in common myeloid and lymphoid progenitors gements all lead to the expression of hybrid that start (CMP and CLP) and except in the erythroid cells, it with the amino-terminal moiety of NUP98. This amino-terminal must be downregulated for the proper differentiation of both domain contains multiple copies of a glycine-leucine-phenyl- lineages to take place.7,8 Hmgb3 RNA has recently been alanine-glycine core motif, known to recruit the CBP/p300 found to be part of an embryonic stem cell-like transcription acetyltransferases.3 Evidence from overexpressions studies in signature that was defined in mouse models of mixed-lineage mouse bone marrow (BM) progenitors indicates that the NUP98- leukemia (MLL)-mediated leukemic transformation and was fusion proteins induce leukemic transformation through the also reported to be transiently upregulated during myeloid upregulation of Hoxa genes and the Hox cofactor Meis1. differentiation.9 Expression of Hoxa, most frequently Hoxa7, Hoxa9 and Hoxa10 We earlier reported the involvement of NUP98 in is postulated to induce a self-renewal stem cell-like program that t(X;11)(q28;p15) in a 73-year-old woman with therapy-related likely contributes to the leukemic process.4,5 acute myeloblastic leukemia (AML) with M4 subtype.1 Fluores- High mobility group (HMG) proteins are non-histone cence in situ hybridization experiments performed on meta- -associated proteins that bind to DNA with limited or phasic of the blast cells permitted the mapping of no sequence specificity. Among them, HMGB proteins the translocation breakpoint on Xq28 in which only the HMGB3 are important architectural facilitators of nucleosome remodel- gene had the correct transcription orientation (telomere to

Leukemia Letters to the Editor 655 ) that would allow an in-frame fusion to NUP98 To establish the transforming properties of the NUP98– (assuming a simple translocation event). Reverse transcriptase- HMGB3 fusion, murine primary BM hematopoietic progenitors PCR using primers located within NUP98 and HMGB3 exons were transduced with a retroviral vector murine stem cell virus showed the presence of NUP98–HMGB3 fusion transcript (MSCV) co-expressing the NUP98–HMGB3 fusion cDNA and from patient’s material and not from control complementary the green fluorescent (GFP), a MSCV–NUP98–HOXA9 DNA (cDNA) (Figure 1a). Nucleotide sequence analyses of the used as a reference or the empty MSCV alone as negative fragment revealed an in-frame fusion of the exon 11 of NUP98 control. Viral supernatants were obtained as described.10 to the exon 2 of HMGB3 (Figure 1b). No reciprocal HMGB3– Transduced cells were split in two and one-half was seeded NUP98 transcript could be detected (not show). The NUP98– in methylcellulose medium for serial replating assays. We HMGB3 predicted protein is composed of the first 422 observed that NUP98–HMGB3-transduced progenitors formed amino acids of NUP98 fused to the entire HMGB3 coding moderate numbers of colonies (90% GFP þ ) until but not sequence, (Figure 1b). Indeed, in the normal HMGB3 tran- beyond the third round of replating (not shown). Cells script, the first coding ATG codon is located in exon 2 after transduced with NUP98–HOXA9 grew beyond five replating, a short untranslated sequence. In the fusion transcript, these whereas for empty vector-transduced cells, no colony was nucleotides accommodate a continuous reading frame from observed after the second replating. This indicates that exon 11 of NUP98 through HMGB3. NUP98–HMGB3 is a weak oncogene in vitro. The other half

NUP98-HMGB3 IR (n=6)

le TTGGTGCAGGATTTGGAACAGTCAGGATGGCTAAAGGTGAC p NUP98-HMGB3 IIR (n=4) m Exon 11 of NUP98 Exon 2 of HMGB3 MSCV (n=3) der 100 t’s sa l lad n A tro tie n RNA a Kb DN p co binding 80 1 3.5 domain 920 NUP98 NH2 GLFG GLFG COOH 2.3 60 1.6 HMG 40 % survival 1 box 200 HMGB3 NH2 COOH 0.7 20

0.3 1 0 NUP98- 624 GLFG GLFG 0 50 100 150 200 HMGB3 NH2 COOH days post BMT

PB BM

X10

X40 MSCV MSCV (n=3) NUP98-HMGB3 IR (n=6) NUP98-HMGB3 IIR (n=4) GFP (%) 10 ± 9 80 ± 7 88 ± 4 WBC (G/L) 8 ± 2 66 ± 91 * 37 ± 17 * Hemoglobin (g/dL) 13 ± 1 6.3 ± 0.6 * 5.2 ± 0.9 * Platelets (G/L) 755 ± 354 99 ± 27 * 163 ± 75 * Spleen weight (g) 0.12 ± 0.03 0.47 ± 0.15 * 0.50 ± 0.07 * NUP98-HMGB3

Figure 1 Molecular breakpoint analysis of the t(X;11)(q28;p15) and bone marrow transplants (BMTs) using NUP98–HMGB3-transduced hematopoietic progenitors. (a) A specific NUP98–HMGB3 product is detected by reverse transcriptase-PCR (RT-PCR) in the patient’s sample at diagnosis. PCR with primers located in NUP98 exon 8 (50-TTGGCCAACAGAATCAGCAGAC-30) and HMGB3 exon 5 (50-CCGGGCAACTTTAG CAGGAC-30) yielded a 983-bp product. Control lane corresponds to the complementary DNA (cDNA) of the human malignant cell line MO7E. (b) Partial nucleotide sequence of the NUP98–HMGB3 chimeric transcript shows that nucleotide 1609 (end of exon 11) of the NUP98 gene and nucleotide 88 of HMGB3 (that is, the start of HMGB3 exon 2) are joined in frame. Schematic representation of the native and chimeric proteins, showing the glycine-leucine-phenylalanine-glycine (GLFG) repeats and RNA binding domain of NUP98 and the two high mobility group (HMG) boxes of HMGB3. (c) Kaplan–Meier survival plot of NUP98–HMGB3 recipients together with control (MSCV) mice. Primary recipients (NUP98– HMGB3 IR, n ¼ 6; MSCV, n ¼ 3) were obtained by the injection of 5 Â 105 transduced Lin– cells into the retroorbital vein of sublethally irradiated mice. Animals were killed because of signs of disease between day 54 and 141 post-transplant. Secondary recipients (NUP98–HMGB3 IIR, n ¼ 4) were obtained by the engraftment of 106 bone marrow (BM) cells from primary mice (killed at day 54 post-transplant) into sublethally irradiated n animals; they all died between day 68 and 87 post-BMT. (d) Hematological parameters of primary and secondary recipients. Po0.05 by Mann– Whitney test. (e) Cytological analysis of peripheral blood (PB) and BM cells, evaluated on May–Gru¨nwald–Giemsa staining of smears and cytospin preparations respectively showed an over-representation of mature myeloid cells in the two tissues, associated with the disappearance of the erythroid and megakaryocytic compartments in the BM for primary NUP98–HMGB3 recipient mice compared with MSCV-transduced animals.

Leukemia Letters to the Editor 656 Gated on GFP+ cells MSCV NUP98-HMGB3 7.94% 58.46% 27.72% 63.53%

Gated on Lin- GFP+ cells

MSCV NUP98-HMGB3

BM MP LSK 25% 4%

1.26% 2.39%

52.01% 61.1%

6.75% 3.88% 26.37% 65.10% c-Kit

Sca-1 spleen

2.87% 0.66%

4.70% 44.84% 23.03% 67.55% GMP 40.12% 90.13%

RII/III CMP

γ 24.50% 4% Fc

MEP 12% 1.70% PB

CD34 CD11b

3.12% 4.17%

Gr-1

1500 NUP98-HMGB3 IR

1200 NUP98-HMGB3 IIR NUP98-HOXA9 900

600

300

80

60

40 Relative expression to MSCV expression Relative

20

0 Hoxa5 Hoxa7 Hoxa9 Hoxa10 Meis1 Pbx1 Pbx3

Figure 2 Cytometric analyses of the degree of hematopoietic proliferation induced by NUP98–HMGB3-fusion and quantification of Hoxa genes expression. (a) Fluorescence-activated cell sorting (FACS) analysis of the bone marrow (BM), spleen and peripheral blood (PB) showed an enrichment for mature Gr-1 þ CD11b þ myeloid cells in primary NUP98–HMGB3-transduced mice, compared with mice engrafted with MSCV- transduced progenitors. (b) FACS analysis of the LinÀSca-1 þ c-Kit þ (LSK) and myeloid progenitor (MP) populations in the BM. MPs from primary NUP98–HMGB3 recipients are enriched in the granulocyte-macrophage progenitor (GMP) subset whereas common myeloid progenitors (CMPs) and megakaryocyte-erythroid progenitors (MEPs) are virtually absent. (c) Real-time reverse transcriptase-PCR (RT-PCR) analysis of endogenous Hoxa (Hoxa5, Hoxa7, Hoxa9, Hoxa10), Meis1, Pbx1 and Pbx3 genes expression in engrafted mice. Accumulation of transcript was quantified in primary (NUP98–HMGB3 IR n ¼ 3) and secondary (NUP98–HMGB3 IIR n ¼ 3) recipients, compared with NUP98–HOXA9 recipients. Levels of expression are standardized to Abl and expressed relative to the expression levels measured in MSCV-engrafted mice. Values shown are mean±s.d. from two independent experiments.

Leukemia Letters to the Editor 657 of NUP98–HMGB3-transduced cells was engrafted into sub- be involved in the leukemogenic properties of NUP98 fusions. lethally irradiated mice. Mice transplanted with progenitors This would be consistent with recent work in mice. Indeed, transduced with the empty MSCV (n ¼ 3) remained free of co-expression of the Hmgb3 Myb, and Cbx5 genes is sufficient hematological disease up to 12 months after transplantation. to induce HoxA/Meis1-independent immortalization of mouse Mice engrafted with progenitors transduced with the NUP98– myeloid progenitors, pointing to a critical role for these genes HOXA9 fusion developed a myeloproliferative disease (MPD) in the leukemic process.9 Thus, we suggest that the ectopic that progressed to AML with long latency (n ¼ 2; median 250 expression of HMGB3, as a result of the fusion with NUP98, can days) consistent with previous reports11 (not shown). Mice that bypass the requirement of a concomitant Hoxa9 and Meis1 received NUP98–HMGB3-transduced cells (n ¼ 6) rapidly died misregulation in human AML. To our knowledge, this is the first with a median survival of 112 days (range, 60–140 days) report of a genetic alteration of an HMGB gene in human (Figure 1c). They developed hyperleucocytosis, anemia with hematological malignancies. It is also the first example of a mucous paleness and dyspnea, thrombopenia and spleno- NUP98 leukemogenic fusion whose expression is not associated megaly (Figure 1d). Femurs of NUP98–HMGB3 mice were with a strong Hoxa and concomitant Meis1 expression. Together, conspicuously discolored reflecting a block in terminal erythro- the data indicate that further investigations of HMGB3-related poiesis, as confirmed by fluorescence-activated cell sorting leukemia may facilitate the analysis of conserved ‘terminal’ (FACS) quantitative analyses performed on BM erythroid mechanisms of AML uncoupled from upstream events. progenitors (not shown). BM smears from femurs showed cytological abnormalities that affected immature and mature myeloid cells, suggesting a MPD-like leukemia according to the Conflict of interest Bethesda classification12 (Figure 1e). In agreement, when compared with control mice, FACS analysis of NUP98–HMGB3 The authors declare no conflict of interest. hematopoietic organs showed an increased number of GFP þ Gr1 þ CD11b þ mature myeloid cells into the peripheral Acknowledgements blood (PB), spleen and BM of sick mice (Figure 2a). BM progenitor populations were analyzed by FACS We thank Julie Bergeron for help in analyses of PB and BM (Figure 2b). The LinÀSca1Àc-Kit þ CD34 þ FcgRII/IIIhigh subset, 13 smears. AP acknowledges the support from the Fondation pour la corresponding to the granulocyte-macrophage progenitors, Recherche Medicale (FRM; Paris, France) and Association pour la was markedly expanded in the BM of NUP98–HMGB3 mice recherche sur le Cancer (ARC; Villejuif, France). This work was compared with control mice. In contrast, the common myeloid funded by grants from ARC and Institut National du Cancer progenitors (CMP) and megakaryocyte-erythroid progenitors (INCa). (MEPs) populations were virtually absent (Figure 2b) indicating that NUP98–HMGB3 expression is associated with preferential A Petit1,2,5, C Ragu1,2, V Della-Valle1,2, MJ Mozziconacci3, expansion of the myelo-monocytic lineages, as observed in M Lafage-Pochitaloff3, G Soler1,2,4, C Schluth1,2,6, several mouse models of MLL-induced leukemias.14 The I Radford1,2,4, C Ottolenghi1,2, OA Bernard1,2, V Penard-Lacronique1,2 and SP Romana1,2,4 malignant nature of NUP98–HMGB3-induced hemopathy was 1 further established by engraftment of primary MPD cells into Institut National de la Sante´ et de la Recherche Me´dicale secondary recipients. All engrafted mice developed an MPD- (Inserm), Hoˆpital Necker-Enfants malades, 149 rue de Se`vres, Paris, F-75015, France; like leukemia of similar immunophenotype (not shown) with 2Universite´ Paris Descartes, 12 rue de l’Ecole de Me´decine, shorter latency than the primary disease (median survival time of Paris, F-75006, France; 80 days) (Figure 1c). Taken together, these results show 3Laboratoire de Cytoge´ne´tique Onco-He´matologique-Hoˆpital that NUP98–HMGB3 acts as an oncogene responsible for Timone Enfants, 264 rue Saint-Pierre, a rapid and transplantable MPD-like leukemia in recipient Marseille, F-13385, France and mice, which is associated with defects in the differentiation of 4Laboratoire de Cytoge´ne´tique, Assistance Publique-Hoˆpitaux myelo-monocytic cells. de Paris (AP-HP), Hoˆpital Necker-Enfants Malades, Mouse models developed to analyze the oncogenic activity of 149 rue de Se`vres, Paris, F-75015, France MLL and NUP98 fusions showed the overexpression of Hoxa9 E-mail: [email protected] or 4,5,15–18 [email protected] and Meis1 in blast cells; in addition, Hoxa9 and Meis1 5Present address: AP-HP, Hoˆpital Armand Trousseau, Service overexpression was associated with the emergence but not the 9 d’He´matologie et d’Oncologie Pe´diatrique, 26 avenue du maintenance of MLL leukemic stem cells. We thus measured Docteur Arnold Netter, Paris, F-75012, France the expression levels of genes of the Hoxa cluster as well as the 6Present address: Service de cytoge´ne´tique, Centre de biologie genes encoding for the HOX cofactors Meis1, Pbx1 and Pbx3 in et de pathologie Est, Hospices civils de Lyon, 59 boulevard BM cells from primary and secondary NUP98–HMGB3 reci- Pinel, Bron, F-69677, France pients. We compared these values to NUP98–HOXA9 leukemic mice. Relative to NUP98–HOXA9, NUP98–HMGB3 BM cells References showed much weaker expression (approximately 10- to 90-fold) of Hoxa5, Hoxa7 and Hoxa9 (Figure 2c). In contrast, roughly 1 Romana SP, Radford-Weiss I, Ben Abdelali R, Schluth C, Petit A, similar transcription of Hoxa10 was observed in all samples, Dastugue N et al. NUP98 rearrangements in hematopoietic whereas Meis1 expression was detected only in NUP98– malignancies: a study of the Groupe Francophone de Cytogene- HOXA9 leukemic cells and Pbx3 transcripts levels were tique Hematologique. Leukemia 2006; 20: 696–706. similarly upregulated in both leukemia models compared with 2 Moore MA, Chung KY, Plasilova M, Schuringa JJ, Shieh JH, Zhou P control. As NUP98–HMGB3-expressing blast cells upregulate et al. NUP98 dysregulation in myeloid leukemogenesis. Ann NY Hoxa9 only weakly and retain wild-type levels of Meis1,we Acad Sci 2007; 1106: 114–142. 3 Kasper LH, Brindle PK, Schnabel CA, Pritchard CE, Cleary ML, infer that transformation mediated by the fusion does not involve van Deursen JM. CREB binding protein interacts with - deregulated activity of the canonical Hoxa–Meis1 pathway. specific FG repeats that activate transcription and mediate NUP98- These results suggest that several transformation pathways might HOXA9 oncogenicity. Mol Cell Biol 1999; 19: 764–776.

Leukemia Letters to the Editor 658 4 Wang GG, Cai L, Pasillas MP, Kamps MP. NUP98-NSD1 links chronic and acute myeloid leukemias in mice. EMBO J 2001; 20: H3K36 methylation to Hox-A gene activation and leukaemo- 350–361. genesis. Nat Cell Biol 2007; 9: 804–812. 12 Kogan SC, Ward JM, Anver MR, Berman JJ, Brayton C, Cardiff RD 5 Wang GG, Song J, Wang Z, Dormann HL, Casadio F, Li H et al. et al. Bethesda proposals for classification of nonlymphoid Haematopoietic malignancies caused by dysregulation of a hematopoietic neoplasms in mice. Blood 2002; 100: 238–245. chromatin-binding PHD finger. Nature 2009; 459: 847–851. 13 Akashi K, Traver D, Miyamoto T, Weissman IL. A clonogenic 6 Travers AA. Priming the nucleosome: a role for HMGB proteins? common myeloid progenitor that gives rise to all myeloid lineages. EMBO Rep 2003; 4: 131–136. Nature 2000; 404: 193–197. 7 Nemeth MJ, Curtis DJ, Kirby MR, Garrett-Beal LJ, Seidel NE, 14 Cozzio A, Passegue E, Ayton PM, Karsunky H, Cleary ML, Cline AP et al. Hmgb3: an HMG-box family member expressed in Weissman IL. Similar MLL-associated leukemias arising from self- primitive hematopoietic cells that inhibits myeloid and B-cell renewing stem cells and short-lived myeloid progenitors. Genes differentiation. Blood 2003; 102: 1298–1306. Dev 2003; 17: 3029–3035. 8 Nemeth MJ, Kirby MR, Bodine DM. Hmgb3 regulates the balance 15 Krivtsov AV, Armstrong SA. MLL translocations, histone modifica- between hematopoietic stem cell self-renewal and differentiation. tions and leukaemia stem-cell development. Nat Rev Cancer 2007; Proc Natl Acad Sci USA 2006; 103: 13783–13788. 7: 823–833. 9 Somervaille TC, Matheny CJ, Spencer GJ, Iwasaki M, Rinn JL, 16 Palmqvist L, Pineault N, Wasslavik C, Humphries RK. Candidate Witten DM et al. Hierarchical maintenance of MLL myeloid genes for expansion and transformation of hematopoietic stem leukemia stem cells employs a transcriptional program shared with cells by NUP98-HOX fusion genes. PLoS ONE 2007; 2: e768. embryonic rather than adult stem cells. Cell Stem Cell 2009; 4: 17 Hirose K, Abramovich C, Argiropoulos B, Humphries RK. 129–140. Leukemogenic properties of NUP98-PMX1 are linked to NUP98 10 Su X, Drabkin H, Clappier E, Morgado E, Busson M, Romana SP and homeodomain sequence functions but not to binding proper- et al. Transforming potential of the T-cell acute lymphoblastic ties of PMX1 to . Oncogene 2008; 27: leukemia-associated genes HOXA13, TLX1, and TLX3. 6056–6067. Genes Cancer 2006; 45: 846–855. 18 Jankovic D, Gorello P, Liu T, Ehret S, La Starza R, Desjobert C et al. 11 Kroon E, Thorsteinsdottir U, Mayotte N, Nakamura T, Sauvageau Leukemogenic mechanisms and targets of a NUP98/HHEX fusion G. NUP98-HOXA9 expression in hemopoietic stem cells induces in acute myeloid leukemia. Blood 2008; 111: 5672–5682.

Nilotinib-mediated inhibition of ABCB1 increases intracellular concentration of dasatinib in CML cells: implications for combination TKI therapy

Leukemia (2010) 24, 658–660; doi:10.1038/leu.2009.242; nilotinib IC50 were 98 and 400 nM, respectively. The combina- published online 10 December 2009 tion indices (CIs) of the dasatinib and nilotinib mixture at effect dose (ED)50, ED75 and ED90 were 0.85, 0.74 and 0.65, We read with interest the recently published article by Davies respectively (Figures 2a and b). These results demonstrate the et al.1 In this study they found nilotinib to be an inhibitor, but synergistic effect of the dasatinib and nilotinib combination in not a substrate of ABCB1 and ABCG2.1 We and others have K562-Dox cells. To assess the cell death mediated by this previously demonstrated that dasatinib is a substrate of combination, K562-Dox cells were cultured with dasatinib, ABCB12,3 and ABCG2.2 Based on our and Davies et al.’s1 nilotinib or the combination (1:4 ratio), and cell death was findings, we speculate that a combination of nilotinib and assessed at 72 h. The combination of dasatinib and nilotinib dasatinib may be synergistic, as nilotinib-mediated inhibition of induced significantly higher cell death than that induced by ABCB1 and ABCG2 may enhance the intracellular uptake and each individual drug alone, and the combination index at ED50 retention (IUR) of dasatinib. This would be analogous to the and ED75 were 0.41 and 0.65, respectively, confirming the earlier described4 synergy between nilotinib and imatinib, synergistic effect of the dasatinib and nilotinib combination in which we speculated to be due to an imatinib-mediated K562-Dox cells (Table 1). These results are consistent with the increase in nilotinib IUR.5 findings of Davies et al.,1 and suggest the potential for synergy Here we report the effect of nilotinib on 14C-dasatinib IUR in between dasatinib and nilotinib in combination. parental and ABCB1-overexpressing cell lines. K562, CCRF- According to Davies et al.,1 nilotinib is also an ABCG2 CEM cells, and their ABCB1-overexpressing respective sublines, inhibitor; however, in our study, nilotinib did not increase K562-Dox and VBL-100, were incubated with 14C-dasatinib dasatinib IUR (19.74±2.86 vs 23.40±2.25; P ¼ 0.1) in K562- with or without nilotinib for 2 h at 37 1C. The aqueous phase and ABCG2 cell line (ABCG2-overexpressing cell line). cell pellet were then separated by centrifugation and the IUR of Davies et al.1 also reported that, although nilotinib inhibited 14C-dasatinib was assessed as described previously.2 Nilotinib ABCB1-mediated efflux in an ABCB1-overexpressing cell line, significantly increased 14C-dasastinib IUR in the ABCB1-over- they could not demonstrate a similar inhibitory effect in chronic expressing K562-Dox (11.47±2.91 vs 34.52±9.78, P ¼ 0.007; myelogenous leukemia (CML)-CD34 þ cells. We compared the Figure 1a) and VBL-100 cells (9.4±2.4 vs 18.24±1.44, effect of nilotinib on 14C-dasatinib IUR in CML-CD34 þ cells Po0.01; Figure 1c), but not in their parental cells. To assess and mononuclear cells (MNCs). Unlike imatinib,6 dasatinib IUR whether increased dasatinib IUR enhances dasatinib potency in in CML-CD34 þ cells was not significantly lower than that in an ABCB1-overexpressing cell line, K562-Dox cells were MNCs (8.22±1.60 vs 8.08±1.12, P ¼ 0.8; n ¼ 6). Moreover, cultured with dasatinib, nilotinib, or a 1:4 ratio (based on nilotinib did not change 14C-dasatinib IUR in MNCs IC50) combination of dasatinib and nilotinib for 2 h, and p-Crkl (8.08±1.12 vs 8.05±1.0, P ¼ 0.9; n ¼ 6) and CML-CD34 þ inhibition was assessed as described earlier.2 The dose effect (7.75±1.69 vs 7.15±2.06, P ¼ 0.8; n ¼ 4) cells. These findings and synergy analysis was performed by using CalcuSyn Software may suggest that, although dasatinib is an ABCB1 substrate, (Biosoft, Cambridge, UK). In K562-Dox cells dasatinib and ABCB1 may not be the major determinant of dasatinib IUR in

Leukemia