Letters to the Editor 1427 mutations is not sufficient to postulate a two-hit model for ALL 2 Marschalek R. Mechanisms of leukemogenesis by MLL fusion proteins. Br J childhood . Therefore, other interpretations are neces- Haematol 2011; 152: 141–154. sary. By now, we can only speculate about a biological function, 3 Bardini M, Spinelli R, Bungaro S, Mangano E, Corral L, Cifola I et al. DNA copy- but there is a significant difference between mutant and number abnormalities do not occur in infant ALL with t(4;11)/MLL-AF4. Leukemia physiological RAS signaling. Downregulation of phosphoinosi- 2010; 24: 169–176. tide-3 kinase signaling and activation the ATM/ATR-induced DNA 4 Mahgoub N, Parker RI, Hosler MR, Close P, Winick NJ, Masterson M et al. RAS damage response system may cause a delay of tumor develop- mutations in pediatric with MLL gene rearrangements. Genes Chro- ment.14 In addition, an additional RAS mutation might be mosomes 1998; 21: 270–275. 5 Taketani T, Taki T, Sugita K, Furuichi Y, Ishii E, Hanada R et al. FLT3 mutations in recognized by the mother’s immune system, and thus the activation loop of tyrosine kinase domain are frequently found in infant ALL preventing tumor outgrowth already in utero. with MLL rearrangements and pediatric ALL with hyperdiploidy. Blood 2004; 103: To this end, further work will be necessary to find a satisfactory 1085–1088. explanation for these observations, and to understand at the 6 Liang DC, Shih LY, Fu JF, Li HY, Wang HI, Hung IJ et al. K-Ras mutations and N-Ras molecular mechanisms triggered by mutated RAS signaling. mutations in childhood acute leukemias with or without mixed-lineage leukemia Understanding the importance of oncogenic signaling for the gene rearrangements. Cancer 2006; 106: 950–956. biology of MLL-rearranged leukemia may bear the potential to 7 Stirewalt DL, Radich JP. The role of FLT3 in haematopoietic malignancies. Nat Rev identify novel drug targets that can be therapeutically addressed Cancer 2003; 3: 650–665. in future therapy regimens. This might be true for signaling events 8 Armstrong SA, Mabon ME, Silverman LB, Li A, Gribben JG, Fox EA et al. FLT3 mutations in childhood acute lymphoblastic leukemia. Blood 2004; 103: 3544–3546. deriving from highly expressed FLT3, as inhibition of this receptor 9 Furuichi Y, Goi K, Inukai T, Sato H, Nemoto A, Takahashi K et al. Fms-like tyrosine by PKC412 and CEP-701 resulted in a selective killing of childhood kinase 3 ligand stimulation induces MLL-rearranged leukemia cells into quies- 15 leukemia cells. cence resistant to antileukemic agents. Cancer Res 2007; 67: 9852–9861. 10 Stam RW, Schneider P, de Lorenzo P, Valsecchi MG, den Boer ML, Pieters R. Prognostic significance of high-level FLT3 expression in MLL-rearranged infant CONFLICT OF INTEREST acute lymphoblastic leukemia. Blood 2007; 110: 2774–2775. The authors declare no conflict of interest. 11 Choudhary C, Olsen JV, Brandts C, Cox J, Reddy PN, Bo¨hmer FD et al. Mislocalized activation of oncogenic RTKs switches downstream signaling outcomes. Mol Cell 2009; 36: 326–339. ACKNOWLEDGEMENTS 12 Ng MH, Ng RK, Kong CT, Jin DY, Chan LC. Activation of Ras-dependent Elk-1 This work is supported by a grant DKS 2011.09 from the German Children Cancer Aid activity by MLL-AF4 family fusion oncoproteins. Exp Hematol 2010; 38: 481–488. to RM, and grant Bu 1854/2-1 from the DFG to AB. We thank OA Haas for critical 13 Moriya K, Suzuki M, Watanabe Y, Takahashi T, Aoki Y, Uchiyama T et al. Devel- discussions. opment of a multi-step leukemogenesis model of MLL-rearranged leukemia using humanized mice. PLoS One 2012; 7: e37892. C Prelle1, A Bursen1, T Dingermann1 and R Marschalek1 14 Mallette FA, Gaumont-Leclerc MF, Ferbeyre G. The DNA damage signaling pathway is 1Institute of Pharmaceutical Biology/DCAL, Goethe-University, a critical mediator of oncogene-induced senescence. Genes Dev 2207; 21:43–48. Frankfurt 60438, Germany 15 Brown P, Levis M, Shurtleff S, Campana D, Downing J, Small D. FLT3 inhibition selectively kills childhood acute lymphoblastic leukemia cells with high levels of E-mail: [email protected] FLT3 expression. Blood 2005; 105: 812–820.

REFERENCES This work is licensed under a Creative Commons Attribution- 1 Meyer C, Kowarz E, Hofmann J, Renneville A, Zuna J, Trka J et al. New insights into NonCommercial-NoDerivs 3.0 Unported License. To view a copy of the MLL recombinome of acute leukemias. Leukemia 2009; 23: 1490–1499. this license, visit http://creativecommons.org/licenses/by-nc-nd/3.0/

Low GFI1 expression in white blood cells of CP–CML patients at diagnosis is strongly associated with subsequent blastic transformation

Leukemia (2013) 27, 1427–1430; doi:10.1038/leu.2013.47 molecular response (MMR; o0.1% BCR–ABL1 (IS)) or early molecular response (BCR–ABL1 p10% following 3 months of imatinib therapy) in de-novo chronic phase–CML (CP–CML) patients has not been examined. We hypothesized, based on the findings of Soliera et al. Chronic myeloid leukemia (CML) is characterized by the BCR–ABL1 that increased GFI1 expression would be associated with a fusion gene, resulting in uncontrolled proliferation of myeloid favorable outcome in CP–CML treated with imatinib. progenitor cells. Growth factor independence 1 (GFI1) is a The expression of GFI1 was examined using TaqMan Low transcription factor with a crucial role in haematopoiesis, including Density array (RQ-PCR, Applied Biosystems, Carlsbad, CA, USA) in preserving haematopoietic stem cell (HSC) quiescence and enhan- white blood cells of 40 de-novo CP–CML patients enrolled in cing granulocytic differentiation, but is not required for inducing clinical trials conducted in our Centre, who received imatinib as myeloid differentiation in p210BCR/ABL-transformed cells.1 frontline therapy. Data analysis was performed using the statistical Recently, Soliera et al.2 demonstrated that ectopic GFI1 expression program R version 2.15.1, with the Bioconductor package inhibited proliferation and colony formation both in p210BCR/ABL- high-throughput analysis and visualization of quantitative real- expressing cell lines and in primary CD34 þ CML cells through the time PCR (HTqPCR)3 and GraphPad Prism 5. GFI1 expression repression of STAT5B and/or Mcl-1. This study, along with their (Hs00382207_m1, Applied Biosystems) was normalized to GUSB previous work1 demonstrated the biological importance of the expression (Hs99999908_m1, Applied Biosystems) using the DDCt GFI1/STAT5B/Mcl-1 regulatory pathway on proliferation and survival method.4 As GFI1 is primarily expressed in ,5 we of CML cells. However, the association between GFI1 expression and rationalized that the use of white cells would enable accurate the achievement of response to imatinib therapy, be it major representation of total GFI1 expression from each patient.

Accepted article preview online 15 February 2013; advance online publication, 8 March 2013

& 2013 Macmillan Publishers Limited Leukemia (2013) 1394 – 1440 Letters to the Editor 1428 p=0.002 p=0.35 p=0.42 0 0 0 -1 -1 -1 -2 -2 -2

Ct, log2) -3 Ct, log2) -3 Ct, log2) -3 Δ Δ Δ (- (- -4 (- -4 -4 GFI1 Expression GFI1 Expression -5 GFI1 Expression -5 -5 Transformed CP-CML No MMR MMR No MMR MMR to BC

p=0.96 p=0.10 0 p=0.55 0 0 -1 -1 -1 -2 -2 -2 Ct, log2) Ct, log2) -3 Ct, log2) -3 -3 Δ Δ Δ (- (- (- -4 -4 -4 GFI1 Expression GFI1 Expression GFI1 Expression -5 -5 -5 ≤10% >10% Low OA High OA Mutation No Mutation

0 p=0.92 -1 -2 -3 Ct, log2) Δ

(- -4

GFI1 Expression -5 Low Medium High Figure 1. Examination of GFI1 expression using CP–CML white cell mRNA. TaqMan-based RQ-PCR was performed for GFI1 expression on white cells from 40 CP–CML patients at diagnosis. The assay was performed in duplicate and normalized to the housekeeping gene GUSB. The average of the duplicate assays was used for data presentation and statistical analysis. The differential expression of GFI1 was examined for: (a) Transformation to BC, (b) MMR outcome by 12 months (including 12 months), (c). MMR outcome by 24 months (including 24 months), (d) 3-month BCR–ABL1 transcript level (IS), (e) OA and (f) KD mutation development during imatinib therapy. (g) Differential GFI1 expression between low, medium and high Sokal score10 groups (one-way analysis of variance , ANOVA). All statistical analysis were performed using the Mann–Whitney U-test, unless otherwise stated. Box plots indicate the interquartile range (25–75%) around the median. Whiskers represent the 10th and 90th percentiles.

We first examined whether there was differential expression of We next examined the expression of GFI1 in CP–CML patients GFI1 at diagnosis, in those patients who transformed to blast crisis who did, or did not, achieve MMR (o0.1% BCR–ABL1 levels) by (BC; time range: 3.7–13 months, n ¼ 6) while receiving imatinib as 12 months. Intriguingly, there was no statistically significant frontline therapy (Figure 1a). A significant decrease in GFI1 difference observed in GFI1 expression (P ¼ 0.35; Figure 1b) expression at diagnosis was observed in patients who transformed between these patients, nor when the response end point was to BC on imatinib therapy, compared with those who did not extended to 24 months (Figure 1c). We then determined whether (n ¼ 37, P ¼ 0.002), suggesting that lower GFI1 expression is GFI1 expression was instead a biomarker for the achievement of associated with the transformation in CML. Furthermore, BCR–ABL1 transcript levels of p10% (IS) at 3 months, a time-point receiver–operator characteristic analysis revealed that de-novo recently demonstrated to be predictive of longer-term response in CP–CML patients with GFI1 DCto 2.782 (92% specificity and imatinib-treated patients.7 However, again there was no 83% sensitivity; P ¼ 0.0036, likelihood ratio ¼ 10.28) are B10 times significant difference observed in GFI1 expression between more likely to transform to BC. To determine whether the patients with 3-month BCR–ABL1 transcript levels p10% (good observed variation in GFI1 expression between patients could response) and those with levels 410% (poor response; Figure 1d). be attributable to differences in cell populations, we examined the The organic cation transporter-1 (OCT-1) activity assay, which correlation between GFI1 expression and differential cell count measures the functional activityoftheOCT-1protein,themajor information available in 32 patients. No significant correlation active transporter involved in imatinib disposition in CML cells, using (P40.05) was observed between the total white cell count, or patient mononuclear cells (MNC) at diagnosis, is also a strong absolute blast, , or numbers, prognostic indicator.8 In particular, patients with very low OCT-1 and GFI1 expression (data not shown). Additionally, no significant activity (OA) (p4 ng/200 000 cells) are at significant risk of poor correlation was observed between GFI1 expression and blast cell molecular response, mutation development and leukemic percentage (data not shown). We also examined the impact of transformation on imatinib therapy.9 To determine whether GFI1 tyrosine-kinase inhibitor treatment on GFI1 expression in vivo by expression was associated with OA in this cohort of CP–CML patients, comparing the expression in blood collected after the first 8 days we grouped patients according to their OA into low (p4 ng/200 000 of imatinib therapy (n ¼ 12; including three patients who later cells) and high (44 ng/200 000 cells) groups. There was no transformed to BC), with that collected from the same patients at statistically significant difference (P ¼ 0.96; Figure 1e) in GFI1 diagnosis. Overall, we observed a significant decrease in GFI1 expression, suggesting that GFI1 expression has no role in OA. expression in all patients after 8 days of imatinib therapy Point mutations in the kinase domain (KD) of BCR–ABL1 are the (Po0.0001), consistent with the previously published results of most prevalent mechanism of acquired resistance to kinase Huang et al.6 Importantly, this decrease was observed in both the inhibitor therapy in CML patients.10 Therefore, the relationship patients who transformed to BC (n ¼ 3, P ¼ 0.02) and those who between mutation development and GFI1 expression in our did not (n ¼ 9, P ¼ 0.0005; data not shown). cohort of patients with mutations (n ¼ 4, within 24 months) was

Leukemia (2013) 1394 – 1440 & 2013 Macmillan Publishers Limited Letters to the Editor 1429 p=0.43 p=0.91 p=0.96 11 11 11 10 10 10 9 9 9 (log2) 8 (log2) 8 (log2) 8 7 7 7 GFI1 Expression GFI1 Expression GFI1 Expression 6 6 6 MMR No MMR MMR No MMR -10% > 10%

p=0.46 p=0.18 p=0.88 11 11 11 10 10 10 9 9 9

(log2) 8 (log2) 8 8 (log2) 7 7 7 GFI1 Expression GFI1 Expression 6 GFI1 Expression 6 6 Low OA High OA Mutation No Mutation Low Medium High

*** n.s. n.s. ** 10 p=0.28 0.5 *** 0.5

8 0.0 0.0 6 -0.5 -0.5 (log2) (log2)

(log2) 4 2 -1.0 -1.0 GFI1 Expression GFI1 Expression GFI1 Expression 0 -1.5 -1.5 NR R CP AP BC CP BC BC Remission Figure 2. Examination of GFI1 expression using CP–CML MNC and CD34 þ mRNA, and at different stages of CML disease course. Microarray data from 93 CP–CML MNC samples was normalized using Robust multiarray average and used to examine the relationship between GFI1 expression with: (a) MMR outcome by 12 months (including 12 months), (b) MMR outcome by 24 months (including 24 months), (c) 3-month BCR–ABL1 transcript level (IS), (d)OA,(e) KD mutation development during imatinib therapy, (f) Sokal score (one-way ANOVA), (g) Imatinib nonresponders (NR) and responders (R) from the McWeeney et al.11 CD34 þ CP–CML microarray data set, (h) Three phases of CML disease progression from the Radich et al.12 microarray data set (*** Po0.001), and (i) CML patients in CP, BC or BC patients who achieved remission after (n ¼ 3; **Po0.01, Mann–Whitney). N.S. indicates not significant (P40.05). All statistical analyses were performed using a two-tailed t-test, unless otherwise stated. Box plots indicate the interquartile range (25–75%) around the median. Whiskers represent the 10th and 90th percentiles. assessed (Figure 1f), however, no significant association was As our preliminary data identified that decreased GFI1 observed. Finally, patients were grouped according to their Sokal expression was associated with CML transformation, we inter- score (low, medium or high), as previously defined,11 but again no rogated the Radich et al.13 microarray data set (GSE4170) to significant difference with respect to GFI1 expression was investigate the expression of GFI1 in CP (n ¼ 42; defined as o10% observed between these groups (Figure 1g). blasts), accelerated phase (AP, n ¼ 17; defined as 10–30% blasts We further examined GFI1 expression in MNC from 93 de-novo or o10% blasts with clonal evolution) and BC–CML (n ¼ 26; CP–CML patients. There was no significant difference in GFI1 defined as 430% blasts). Interestingly, the material from CML expression between patients who achieved MMR by 12 months patients in BC had significantly lower GFI1 expression compared (P ¼ 0.43;Figure2a),orby24months(P ¼ 0.91; Figure 2b), and with CP–CML (Po0.001; Figure 2h) and AP–CML (Po0.001; those who didn’t. Again, as seen in the analysis of white cells, no Figure 2h), supporting our observation that decreased GFI1 significant association between GFI1 expression and the achieve- expression is associated with progression in CML. When we ment of 3-month BCR–ABL1 transcript levels (P ¼ 0.96; Figure 2c), OA compared the GFI1 expression of patients in BC versus BC patients (P ¼ 0.46; Figure 2d), BCR–ABL1 KD mutation development (P ¼ 0.18; who were now in remission after chemotherapy (non-matched Figure 2e) or Sokal score (P ¼ 0.88; Figure 2f) was observed. samples, BC-rem; n ¼ 3), we observed a significant increase in GFI1 As GFI1 has a role in HSC maintenance, and the biological expression in the BC-rem patients (P ¼ 0.009; Figure 2i) to levels importance of GFI1 was originally determined in CD34 þ cells by comparable with CP–CML (P ¼ 0.33; Figure 2i). Furthermore, these Soliera et al.,2 we examined GFI1 expression in CD34 þ cells from data are in keeping with the lower GFI1 expression we observed in CP–CML patients. To do this, the McWeeney et al.12 (GSE14671) our cohort of de-novo CP–CML patients who later transformed to microarray data set involving CD34 þ cells from 59 primary CP– BC. However, it must be noted that an independent study of larger CML patients originally generated to identify genes associated sample size is required to confirm our observation that low GFI1 with imatinib response was reanalysed for GFI1 expression. There expression at diagnosis may be associated with transformation was no significant difference in GFI1 expression between imatinib and progression to BC in CML. responders (patients who achieved at least partial cytogenetic This work supports the previously described role of GFI1 in response within 12 months of therapy) and nonresponders (all inhibition of proliferation and colony formation of p210BCR/ABL- other patients) in CP–CML CD34 þ cells (P ¼ 0.28; Figure 2g). transformed cells and primary CD34 þ CML cells;2 blockade of Taken together, these data suggest that GFI1 expression levels are granulopoiesis in GFI1 / myeloid progenitors via elevated not significantly altered in cases of imatinib resistance, regardless levels of HoxA9, Pbx1 and Meis1;14 and the development of of the cell type examined. myelo-proliferative disease15 and defects to T-cell development16

& 2013 Macmillan Publishers Limited Leukemia (2013) 1394 – 1440 Letters to the Editor 1430 in GFI1 / mice. Taken together, our results suggest that unlike 3 Dvinge H, Bertone P. HTqPCR: high-throughput analysis and visualization of other negative prognostic markers (such as OA, BCR–ABL1 quantitative real-time PCR data in R. Bioinformatics 2009; 25: 3325–3326. transcript levels at 3 months and Sokal score), low GFI1 4 Livak KJ, Schmittgen TD. Analysis of relative gene expression data using expression correlates selectively with a high risk of early real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods 2001; 25: transformation to BC, but not mutation development, primary 402–408. resistance or secondary resistance in CP–CML patients. 5 van der Meer LT, Jansen JH, van der Reijden BA. Gfi1 and Gfi1b: key regulators of hematopoiesis. Leukemia 2010; 24: 1834–1843. 6 Huang M, Hu Z, Chang W, Ou D, Zhou J, Zhang Y. The growth factor indepen- dence-1 (Gfi1) is overexpressed in chronic myelogenous leukemia. Acta Haematol CONFLICT OF INTEREST 2010; 123: 1–5. The authors declare no conflict of interest. 7 Marin D, Ibrahim AR, Lucas C, Gerrard G, Wang L, Szydlo RM et al. Assessment of BCR-ABL1 transcript levels at 3 months is the only requirement for predicting outcome for patients with chronic myeloid leukemia treated with tyrosine kinase ACKNOWLEDGEMENTS inhibitors. J Clin Oncol. 2012; 30: 232–238. 8 White DL, Saunders VA, Quinn SR, Manley PW, Hughes TP. Imatinib increases the DBW is supported by a PhD scholarship from the Leukemia Foundation of Australia. intracellular concentration of nilotinib, which may explain the observed synergy TPH is an NHMRC Practitioner Fellow. This project was funded by a Leukemia and between these drugs. Blood 2007; 109: 3609–3610. Society (LLS) grant and a NHMRC project grant. 9 White DL, Dang P, Engler J, Frede A, Zrim S, Osborn M et al. Functional activity of the OCT-1 protein is predictive of long-term outcome in patients with chronic- CH Kok1,2,5, DB Watkins1,2,5, T Leclercq1, RJ D’Andrea1,2,3, phase chronic myeloid leukemia treated with imatinib. J Clin Oncol. 2010; 28: TP Hughes1,2 and DL White1,2,4 2761–2767. 1Haematology Department, SA Pathology (RAH Campus), Centre of 10 O’Hare T, Eide CA, Deininger MW. Bcr-Abl kinase domain mutations, drug resis- Cancer Biology, Adelaide, South Australia, Australia; tance, and the road to a cure for chronic myeloid leukemia. Blood 2007; 110: 2 2242–2249. Department of Medicine, Centre for Personalised Cancer Medicine, 11 Sokal JE, Cox EB, Baccarani M, Tura S, Gomez GA, Robertson JE et al. Prognostic University of Adelaide, Adelaide, South Australia, Australia; discrimination in ‘good-risk’ chronic granulocytic leukemia. Blood 1984; 63: 789–799. 3 Department of Haematology and Oncology, The Queen Elizabeth 12 McWeeney SK, Pemberton LC, Loriaux MM, Vartanian K, Willis SG, Yochum G et al. Hospital, Woodville, South Australia, Australia and A gene expression signature of CD34 þ cells to predict major cytogenetic 4Faculty of Health Science, University of South Australia, Adelaide, response in chronic-phase chronic myeloid leukemia patients treated with ima- South Australia, Australia tinib. Blood 2010; 115: 315–325. E-mail: [email protected] 13 Radich JP, Dai H, Mao M, Oehler V, Schelter J, Druker B et al. Gene expression 5This authors contributed equally to this work. changes associated with progression and response in chronic myeloid leukemia. Proc Natl Acad Sci USA 2006; 103: 2794–2799. 14 Horman SR, Velu CS, Chaubey A, Bourdeau T, Zhu J, Paul WE et al. Gfi1 integrates progenitor versus granulocytic transcriptional programming. Blood 2009; 113: REFERENCES 5466–5475. 1 Lidonnici MR, Audia A, Soliera AR, Prisco M, Ferrari-Amorotti G, Waldron T et al. 15 Khandanpour C, Kosan C, Gaudreau MC, Duhrsen U, Hebert J, Zeng H et al. Expression of the transcriptional repressor Gfi-1 is regulated by C/EBPa and is Growth factor independence 1 protects hematopoietic stem cells against apop- involved in its proliferation and colony formation-inhibitory effects in p210BCR/ tosis but also prevents the development of a myeloproliferative-like disease. ABL-expressing cells. Cancer Res 2010; 70: 7949–7959. Stem Cells 2011; 29: 376–385. 2 Soliera AR, Mariani SA, Audia A, Lidonnici MR, Addya S, Ferrari-Amorotti G et al. Gfi-1 16 Yucel R, Karsunky H, Klein-Hitpass L, Moroy T. The transcriptional repressor Gfi1 inhibits proliferation and colony formation of p210BCR/ABL-expressing cells via affects development of early, uncommitted c-Kit þ progenitors and transcriptional repression of STAT 5 and Mcl-1. Leukemia 2012; 26: 1555–1563. CD4/CD8 lineage decision in the thymus. J Exp Med 2003; 197: 831–844.

Increased circulating IL-2Ra (CD25) predicts poor outcome in both indolent and aggressive forms of : a comprehensive cytokine–phenotype study

Leukemia (2013) 27, 1430–1433; doi:10.1038/leu.2013.11 (that is aggressive SM (ASM), SM associated with clonal non-mast cell lineage disease (SM-AHNMD) or ). In contrast, patients with indolent SM (ISM) have a normal life expectancy; their risk of transformation to acute leukemia or ASM has been estimated to be 1% and 3%, respectively.3,4 Predictors of The clinical phenotype of systemic mastocytosis (SM) is highly disease progression include advanced age, increased serum variable; establishing prognosis in individual patients is not trivial. b2-microglobulin level and multilineage presence of KITD816V.3 The classification of adult SM patients per the 2008 World Health Recent studies in primary myelofibrosis,5 myelodysplastic syndr- Organization (WHO) system1 has been validated for its prognostic ome6 and diffuse large B-cell lymphoma7 have demonstrated an utility.2 In addition to WHO SM subtype, an independent abnormal circulating cytokine milieu reflecting the host response to association between inferior survival and advanced age, weight clonal proliferation. In these studies, specific cytokines were found to loss, , thrombocytopenia, hypoalbuminemia and excess be prognostically relevant, independent of known disease-specific (BM) blasts has been demonstrated.2 Aside from prognostic factors. Here, we conducted a comprehensive analysis of advanced age, however, the aforementioned parameters are circulating cytokines/chemokines with clinicopathologic and clinical typically applicable only to patients with advanced SM subtypes outcome correlations in a cohort of SM patients.

Accepted article preview online 15 January 2013; advance online publication, 1 February 2013

Leukemia (2013) 1394 – 1440 & 2013 Macmillan Publishers Limited