Search

The genetics of classical Hodgkin lymphoma (cHL) is poorly understood. The finding of a JAK2-involving t(4;9)(q21;p24) in 1 case of cHL prompted us to characterize this translocation on a molecular level and to determine the prevalence of JAK2 rearrangements in cHL. We showed that the t(4;9)(q21;p24) leads to a novel SEC31A-JAK2 fusion. Screening of 131 cHL cases identified 1 additional case with SEC31A-JAK2 and 2 additional cases with rearrangements involving JAK2. We demonstrated that SEC31A-JAK2 is oncogenic in vitro and acts as a constitutively activated tyrosine kinase that is sensitive to JAK inhibitors. In vivo, SEC31A-JAK2 was found to induce a T-lymphoblastic lymphoma or myeloid phenotype in a murine bone marrow transplantation model. Altogether, we identified SEC31A-JAK2 as a chromosomal aberration characteristic for cHL and provide evidence that JAK2 rearrangements occur in a minority of cHL cases. Given the proven oncogenic potential of this novel fusion, our studies provide new insights into the pathogenesis of cHL and indicate that in at least some cases, constitutive activation of the JAK/STAT pathway is caused by JAK2 rearrangements. The finding that SEC31A-JAK2 responds to JAK inhibitors indicates that patients with cHL and JAK2 rearrangements may benefit from targeted therapies.

The genetics of classical Hodgkin lymphoma (cHL) is poorly understood. The finding of a JAK2-involving t(4;9)(q21;p24) in 1 case of cHL prompted us to characterize this translocation on a molecular level and to determine the prevalence of JAK2rearrangements in cHL. We showed that the t(4;9)(q21;p24) leads to a novel SEC31A-JAK2fusion. Screening of 131 cHL cases identified 1 additional case with SEC31A-JAK2and 2 additional cases with rearrangements involving JAK2. We demonstrated that SEC31A-JAK2 is oncogenic in vitro and acts as a constitutively activated tyrosine kinase that is sensitive to JAK inhibitors. In vivo, SEC31A-JAK2 was found to induce a T-lymphoblastic lymphoma or myeloid phenotype in a murine bone marrow transplantation model. Altogether, we identified SEC31A-JAK2as a chromosomal aberration characteristic for cHL and provide evidence that JAK2rearrangements occur in a minority of cHL cases. Given the proven oncogenic potential of this novel fusion, our studies provide new insights into the pathogenesis of cHL and indicate that in at least some cases, constitutive activation of the JAK/STAT pathway is caused by JAK2rearrangements. The finding that SEC31A-JAK2 responds to JAK inhibitors indicates that patients with cHL and JAK2rearrangements may benefit from targeted therapies.

Genetic aberrations of the 9p terminal region are recurrent in hematological malignancies. Many of them involve the JAK2 gene located at 9p24, which is non-randomly targeted by chromosomal translocations, amplifications and mutations in various types of leukemia, lymphoma and myeloproliferative neoplasms. Of interest, recent reports showed that also PDL1 and PDL2, two genes located 350kb centromeric to JAK2, are involved in 9p24 aberrations. These genes, coding for the programmed death ligands, were identified as key targets of 9p24 amplification in primary mediastinal B-cell lymphoma (PMBCL) and classical Hodgkin lymphoma (cHL). Moreover, CIITA-PDL1 and CIITA-PDL2 fusions, leading to overexpression of PDL1 and PDL2, respectively, have been reported in two PMBCL cases.The objective of this study is to characterize structural and numerical 9p24 aberrations in lymphoid malignancies and to assess the molecular consequences of these abnormalities.Eighty-four cases of B-cell non-Hodgkin lymphoma (B-NHL) and 4 cases of T-cell NHL with 9p24 aberrations were selected for this study. In addition, we included 130 cases of cHL with unknown 9p24 status. All cases were screened by FISH with a dual color break apart assay for JAK2 containing probes covering PDL1 and PDL2. Cases with rearrangements of these loci were further characterized by FISH with BAC and fosmid probes, and by aCGH. In available cases, qRT-PCR and/or IHC was used to determine the expression level of candidate genes.Structural aberrations of 9p24 were identified in 8 patients presenting with CLL in Richter transformation (1), MALT/MZL (2), FL (1), DLBCL (2) and ALL/LL (2). The 9p24 breakpoints were further mapped to a 200 kb region centromeric to JAK2 and harboring PDL1 and PDL2. These aberrations (translocations, insertion, inversion) involved either immunoglobulin (IG) genes (3 cases), or non-IG sequences at 1p34/PHACTR4, 4p13, 3p25, 6q13/EEF1A1 and 13q11 (5 cases, including one with a three-way translocation also involving IGH). In one case with inv(9)(p24p23), a microdeletion affecting the 3' region of JAK2 and the 5' region of PDL1 was detected by FISH and aCGH. The expression pattern of 5 candidate genes analyzed in 5 available cases showed upregulation of PDL1 and PDL2 in 1 case, overexpression of sole PDL2 in 2 cases and a normal level of analyzed transcripts in 2 cases. In both cases with t(9;14)(p24;q32) translocation of JAK2 to der(14) was demonstrated by FISH. Unfortunately, these cases were not available for qRT-PCR.9p24 amplification hallmarked by homogeneously staining region (hsr) was identified in 10 cases of PMBCL and 4 cases of T-NHL. FISH screening of 130 cases of cHL identified 54 lymphomas with low copy number gains (5–10 copies) and 30 cases with 9p24 amplification, including 6 cases with a predominant amplification of the region covering PDL1 and PDL2. In addition, putative involvement of PDL1/PDL2 in unbalanced translocations associated with loss of JAK2 and telomeric 9p24 sequences was found in 3 cases of cHL.Expression of the JAK2, PDL1 and PDL2 mRNA was analyzed in 7 cases of PMBCL and 2 cases of T-NHL. Among the PMBCL samples, 4 showed overexpression of all 3 transcripts, 2 showed upregulation of PDL1 and PDL2, and one showed overexpression of sole PDL1. JAK2 and PDL1 were found to be highly expressed in one case of T-NHL, while a case of ALK+ ALCL revealed only upregulation of PDL1.IHC analysis of P-JAK2, PDL1 and PDL2 proteins in available cases is ongoing.Our study demonstrates that the 9p24 region is recurrently targeted by structural and numerical aberrations in B- and T-cell leukemia/lymphoma. These abnormalities may represent primary as well as secondary genetic changes acquired during progression of the disease. Structural aberrations predominantly involve PDL1 and/or PDL2 that seem to be deregulated by their juxtaposition with regulatory sequences provided by partner genes. Five postulated PDL1/PDL2 partner genes include IGH/14q32, IGL/22q11, PHACTR4/1p34.3 EEF1A1/6q13 and JAK2/9p24. In the analyzed PMBCL and T-NHL cases, 9p24 amplification constantly targets PDL1 and is randomly associated with an aberrant co-expression of either JAK2 or PDL2. Given the role of PDL1/PDL2 in inhibition of T-cell activation, their impaired expression in B- and T-cell lymphoma may lead to “T-cell exhaustion” and deficient cellular immunity favoring tumor progression.No relevant conflicts of interest to declare.

Abstract 2460

No relevant conflicts of interest to declare.

T-cell lymphoblastic leukemia (T-ALL) arises from clonal expansion of a lymphoid progenitor that has undergone stepwise alteration at distinct stages of differentiation. It is suggested that a set of cooperative mutations that affect different pathways are required before thymocytes become fully malignant. Despite major improvements in our understanding of the molecular genetics of T-ALL, the underlying mechanisms that lead to the abnormal proliferation and enhanced survival of the leukemic cells remain largely unknown.Array CGH analysis revealed an acquired homozygous microdeletion at chromosome 18p11 in 6 % of T-ALL cases. The deleted region was only 125 kb in size and restricted to the PTPN2 (protein tyrosine phosphatase, non-receptor type 2) locus. PTPN2 encodes an intracellular non-transmembrane tyrosine-specific phosphatase that functions as a negative regulator of a variety of signaling proteins including several members of the janus kinase (JAK) and of signal transducer and activator of transcription (STAT) families, growth factor receptors and SRC family kinases. Homozygous deletion of PTPN2 was specifically found in cases with aberrant expression of the TLX1 transcription factor, with two cases also harboring the NUP214-ABL1 fusion. Analysis of additional TLX1 positive cases by quantitative PCR identified loss of one copy of PTPN2 in 5 out of 20 cases. No mutations were detected in the coding region of PTPN2. To determine the effect of loss of PTPN2 in T-cells, we downregulated the expression of PTPN2 using RNAi technology. siRNA mediated knock-down of PTPN2 affected activation of JAK1 associated cytokine receptors implicated in T-cell development. Ligand stimulation of IL7 and interferon gamma receptor resulted in an augmented and prolonged phosphorylation of JAK1 as well as downstream targets STAT1 and STAT5 in T-ALL cell lines with knock-down of PTPN2. In addition, knock- down of Ptpn2 sensitized the pro B-cell line Ba/F3 to transformation by wild type JAK1 confirming a clear relationship between loss of PTPN2 and JAK1 activation. Knock-down of PTPN2 expression also provided a proliferative advantage and reduced sensitivity to kinase inhibitors in lymphoblastic leukemia cell lines HSB-2 and ALL-SIL.In conclusion, our data provide genetic and functional evidence for a tumor suppressor role of PTPN2 in T-ALL and warrant testing of JAK inhibitors for the treatment of this specific subset of T-ALLs as well as further analysis of a potential negative impact of loss of PTPN2 on responsiveness to anti-cancer treatments.Ferrando: Merck, Pfizer: Research Funding.

No relevant conflicts of interest to declare.

Abstract 141

Abstract 140This icon denotes an abstract that is clinically relevant.

Abstract 2402

Activated tyrosine kinases are implicated in the pathogenesis of chronic and acute leukemia, and represent attractive targets for therapy. Sorafenib (BAY43-9006, Nexavar®) is a small molecule B-RAF inhibitor that is used for the treatment of renal cell carcinoma. Sorafenib also has activity against receptor tyrosine kinases from the PDGFR and VEGFR families. Here we show that sorafenib is a potent inhibitor of the ETV6-PDGFRβ tyrosine kinase that is associated with chronic myelomonocytic leukemia. Oncogenic FLT3 receptor tyrosine kinase mutants implicated in the pathogenesis of acute myeloid leukemia are also potently inhibited by sorafenib, including some of the resistant mutants that confer resistance to PKC412 and other FLT3 inhibitors. Sorafenib induced a cell cycle block and apoptosis in the AML cell lines MV4-11 and MOLM-13, both expressing FLT3 with an internal tandem duplication. In contrast, the imatinib resistant KIT(D816V) mutant that is associated with systemic mastocytosis was found to be resistant to sorafenib. These results suggest that sorafenib could be used for the treatment of myeloid leukemias expressing activated forms of PDGFRβ, FLT3, and some PKC412 resistant mutants, but not for leukemias expressing imatinib resistant KIT(D816V).

Background Molecular analysis of T-cell acute lymphoblastic leukemia (T-ALL) has provided evidence that a stepwise alteration of at least four specific pathways is required during transformation of thymocytes to leukemic T-cells. Genetic alterations in hematopoietic precursor cells lead to loss of cell cycle control, impaired differentiation, proliferation and survival advantages, and unlimited self-renewal capacity. These defects include inactivation of CDKN2A (P16) present in 96 % of the patients, deregulated expression of transcription factors, and mutation of NOTCH1 in 56% of patients. However, the molecular lesions leading to the proliferative and survival advantages of T-ALL cells are less well characterized, remaining unknown in 80 % of the T-ALLs. Aims Our aim was to set up a genome-wide analysis of T-ALL in order to detect cryptic deletions and amplifications, with a special focus on the 90 protein tyrosine kinase genes present in the human genome. Methods We used the array-CGH (micro-array comparative genomic hybridization) technology with slides containing genomic BAC probes spaced every 1 Mb over the human genome. An additional 480 probes were added covering the genomic locations of each of the 90 protein tyrosine kinases genes. A total of 27 T-ALL cases and 12 cell lines were included in the study. Results An interstitial deletion on chromosome 9p24 directly upstream of JAK2 was identified in one patient. The deletion was confirmed by FISH. Quantitative PCR (qPCR) analyses indicated that the deletion was 700 kb in size including exons 1–4 of JAK2. In two cell lines, deletions affecting ALK and ERBB4 were detected. Molecular analyses to characterize the possible presence of fusion transcripts involving tyrosine kinases are in progress. We did not detected other rearrangements involving tyrosine kinase genes in neither of the 26 other T-ALL cases nor in the 10 cell lines, indicating that cryptic deletions or amplifications involving tyrosine kinase genes are relatively rare in T-ALL. The most frequent aberration was the deletion of CDKN2A (16/27 cases and 9/12 cell lines) ranging from only one clone to almost the complete 9p arm. MYB duplication was found in two cases and 4 cell lines, and was confirmed by qPCR and FISH analysis. Two of the 4 cell lines had overexpression of MYB detected by qPCR, which can interfere with apoptosis enhancing the survival of the cells, as has been previously described. PTEN deletion was present in one case and one cell line. Other unbalanced aberrations of various size were detected: del (2p) in 5/12 cell lines, del (5q) in 1/27 samples and 4/12 cell lines, del(6q) in 3/27 samples and 2/12 cell lines, or del(9p) in 5/27 samples and 4/12 cell lines. Some of these rearrangements were not observed by standard cytogenetics. Strikingly, cell lines had significantly more chromosomal abnormalities than primary T-ALL cases. Conclusions We detected a novel cryptic rearrangement of JAK2 in one T-ALL case, and duplication of MYB in two T-ALL cases (2/27; 7.5%) and four cell lines (4/12; 33%). Our results suggest that cryptic deletions or amplifications involving tyrosine kinase genes are relatively rare in T-ALL.

Background

Activated tyrosine kinases are implicated in the pathogenesis of chronic and acute leukemia, and represent attractive targets for therapy. Sorafenib (BAY43-9006, Nexavar®) is a small molecule B-RAF inhibitor that is used for the treatment of renal cell carcinoma. Sorafenib also has activity against receptor tyrosine kinases from the PDGFR and VEGFR families. Here we show that sorafenib is a potent inhibitor of the ETV6-PDGFRβ tyrosine kinase that is associated with chronic myelomonocytic leukemia. Oncogenic FLT3 receptor tyrosine kinase mutants implicated in the pathogenesis of acute myeloid leukemia are also potently inhibited by sorafenib, including some of the resistant mutants that confer resistance to PKC412 and other FLT3 inhibitors. Sorafenib induced a cell cycle block and apoptosis in the AML cell lines MV4-11 and MOLM-13, both expressing FLT3 with an internal tandem duplication. In contrast, the imatinib resistant KIT(D816V) mutant that is associated with systemic mastocytosis was found to be resistant to sorafenib. These results suggest that sorafenib could be used for the treatment of myeloid leukemias expressing activated forms of PDGFRβ, FLT3, and some PKC412 resistant mutants, but not for leukemias expressing imatinib resistant KIT(D816V).

Patients with 3q21q26 rearrangements seem to share similar clinicopathologic features and a common molecular mechanism, leading to MDS or AML. The overexpression of EVI1 (3q26) has been implicated in the dysplasia that characterizes this subset of myeloid neoplasias with poor prognosis. However, EVI1 overexpression had been detected in 9% of AML cases without 3q26 abnormalities; and several 3q21q26 cases with no EVI1 overexpression had been reported. Our aim was to characterize different subgroups in 3q rearrangements, and to define the gene expression profile of these groups, specially in the group with 3q21q26 rearrangements. We analyzed 44 samples (6 cell lines and 38 samples of patients) with 3q rearrangements at G-band level, 36 had AML and 8 MDS. As a control, we analyzed 6 samples of patients with 3q rearrangements and lymphoid neoplasias, a cDNA panel of normal human tissues (Clontech), and 10 normal bone marrow samples. FISH analysis was performed using six BACs: PR11-390G14, RP11-475N22 (GATA2), RP11-689D3 (RPN1) on 3q21; RP11-82C9 (EVI1), RP11-115B16 (MDS1) and RP11-196F13 (TRAIL) on 3q26; and a centromere probe. Expression of GATA2, MDS1-EVI1, and EVI1 was measured by a quantitative real-time RT-PCR (Taqman) assay (Applied Biosystems). Samples were first classified attending at the G-banding. FISH analysis showed the heterogeneity of breakpoints in these cases, and the samples were grouped as 3q21 (5 cases), 3q26 (11 cases), 3q21q26 (16 cases) and other 3q rearrangements (12 cases). Real time RT-PCR showed overexpression (OE) of EVI1 and/or GATA2 in 75% of all the cases (33/44), suggesting the important role of these genes in patients with 3q rearrangements and myeloid neoplasias. In patients with lymphoid neoplasias and 3q rearrangements no genes were found to be OE. A different expression profile was found in the groups analyzed (Table 1). In the 3q21 group, 40% of the patients had GATA2 OE, but no EVI1 or MDS1-EVI1 OE was found. Our data show that EVI1 OE could be specific of the groups 3q26 (64%) and 3q21q26 (69%). The number of cases with EVI1 OE was significantly higher in patients with 3q rearrangements than in those with complex karyotype (data not show). In the group with 3q21q26, EVI1 and/or GATA2 were overexpressed in 87.5% of the cases (14/16). In conclusion, we have detected a different expression profile of the genes analyzed among samples with different 3q rearrangements. Our results suggest a more complex mechanism in patients with 3q21q26 rearrangements, in which GATA2 overexpression may contribute with EVI1 to the leukemogenic progress. Further molecular studies are in progress to analyze the different EVI1 transcripts. Table 1: Incidence of the overexpression of EVI1, GATA2 and MDS1-EVI1 in 32 patients with 3q21 and/or 3q26 rearrangements. Frequency & Percentage EVI1 GATA2 MDS1-EVI1 3q21 0% (0/5) 40% (2/5) 0% (0/5) 3q26 64% (7/11) 73% (8/11) 36% (4/11) 3q21q26 75% (12/16) 69% (11/16) 13% (2/16)

The presence of recurrent cytogenetic aberrations in AML has diagnostic and prognostic value; however, more than 40% of patients have a normal karyotype. Although the presence of FLT3 mutations, or overexpression (OE) of WT1, EVI1 or BAALC, have been reported to discriminate in this group those with a worse prognosis, more molecular markers are necessary to achieve treatment stratification in this heterogeneous group. The mice homolog of EVI1 (3q26) has been reported to have a role in the HSC proliferation through Gata2 expression, a downstream target gene of the Evi1 oncogene. Therefore, GATA2 (3q21), a transcription factor with a relevant role in hematopoiesis, could be a candidate gene in the leukemogenic transformation of AML. Our aim was to determinate the incidence of GATA2 OE in AML, and its association with other well-defined prognostic factors. We analyzed 7 cell lines and 192 samples of AML patients at diagnosis: 61 were in the favorable prognosis group, 82 in the intermediate (68 with normal karyotype) and 56 in the unfavorable (27 with 3q rearrangements). Expression of GATA2, MDS1/EVI1, EVI1 and WT1 were measured by a RTQ-PCR Taqman assay (Applied Biosystems). Mutations of FLT3 (ITD and D835) were also analyzed. Gene expression arrays were performed with the Human 19K Oligo Array (Center for Applied Genomics, University of New Jersey). We detected a different expression profile among the 3 prognosis groups: 42% and 12% of all AML cases had GATA2 and EVI1 OE respectively, and the frequency increased in the poorer prognosis groups (Table 1). GATA2 OE was detected in 46% of cases with normal karyotype, and was more frequent among samples with FLT3-ITD (68% vs 34%; p=0.0021). Among AML-M1 cases, the number of samples with FLT3-ITD and GATA2 OE was twice than in samples without FLT3 mutations (78% vs 39%). All patients with AML-M1, FLT3-ITD and GATA2 OE had WT1 OE. Gene expression array technology was performed to compare 3 groups of AML-M1: I) FLT3 no mutated and GATA2 OE; II) FLT3 no mutated and GATA2 no OE; and III) FLT3-ITD and GATA2 OE. Unsupervised analysis clearly classified these 3 groups, suggesting they could be prognostically relevant AML subgroups, and allowing the detection of 38 candidate genes that were differentially expressed. Our results show that GATA2 OE is a common event in AML cases with normal karyotype (46%). The mutational pattern FLT3-ITD/GATA2/WT1 could define a subgroup of patients with normal karyotype and AML-M1, with a different expression pattern. Incidence of the OE of GATA2 and EVI1 in 199 patients with AML GATA2 EVI1 Frequency p value Frequency p value FG: favorable group IG: intermediate group UG: unfavorable group ns: no significative. FG vs IG 26% (16/61) vs 40% (33/82) ns 0% (0/61) vs 4% (3/82) ns FG vs UG 26% (16/61) vs 57% (32/56) 0.01 0% (0/61) vs 38% (21/56) 0.001 IG vs UG 40 (33/82) vs 57% (32/56) ns 4% (3/82) vs 38% (21/56) 0.001

AML is the consequence of collaboration between at least two classes of mutations: class I, exemplified by constitutively activated tyrosine kinases or their downstream effectors, which confer a proliferative and/or survival advantage to hematopoietic cells; and class II mutations, that result in loss of function of transcription factors that are important for normal hematopoietic differentiation. The activating V617F mutation of JAK2 tyrosine kinase has been recently described as a common event in MPD. This mutation seems to disrupt the JH2 inhibitory regulation of JAK2, leaving the enzyme constitutively active. The same mutation was also found in a small number of patients with either atypical MPD or MDS. However, its incidence in AML is unclear. We investigated the incidence of the tyrosine kinase JAK2 mutation V617F in patients with AML, and its association with other factors with a prognosis meaning in this disease. We focused in cases with normal karyotype, which account for more than 40% of all AML, and that represent a therapeutic challenge because of lack of established therapy protocols. Material and methods. We analyzed 7 cell lines and 218 samples of AML patients at diagnosis: 38 included in the favorable prognosis group, 143 in the intermediate (137 with normal karyotype), and 45 in the unfavorable. DNA was extracted by standard procedures from bone marrow. V617F genotyping was performed by amplification refractory mutation system (ARMS) as previously described (Jones el al., 2005). Mutations of FLT3 (ITD and D835) were also analyzed. Results. We found the mutation V617F in 4.5% of overall AML cases (10/225), and in 4.4% of AML patients with normal karyotype (6/137). A high prevalence was observed in patients with the M2 subtype (7/61, 11.5%): 5 cases had a normal karyotype, one t(8;21), and another inv(16). Taking into account cases with AML-M2 and normal karyotype, the incidence of the V617F mutation was 19% (5/26). The mutation was heterozygous in most cases; we only detected homozygous mutation in 2 cell lines with AML-M6 and M7, both presenting a complex karyotype and 3q rearrangements. One of the cell lines had an addition on 9p24, supporting previous data that imply that 9pLOH, caused by mitotic recombination, could be responsible for the transition from heterozygous to homozygous V617F. Interestingly, one of the cases with M2 had both FLT3 and JAK2 mutations. FLT3 mutations were no detected in the other cases. Genetic evidence in vivo and studies on proliferation and phosphorylation in vitro indicate that the V617F mutation confers a proliferative and survival advantage. Patients carrying the V617F mutation had significantly longer disease duration and more complications in MPD, suggesting a possible prognosis meaning in AML patients. In conclusion, we have identified an acquired JAK2 mutation, which induces a constitutive JAK2 signaling, in 19% of patients with AML-M2 and a normal karyotype. Our data suggest that this activating JAK2 mutation may play an important role in the leukemogenesis mechanism of AML, and could contribute to define a subgroup within patients with AML and normal karyotype. Nevertheless, the presence of the V617F mutation opens the possibility for new targeted therapies in AML. Further studies of the clinical significance of the JAK2 mutation in AML-M2 will be of interest.

AML is the consequence of collaboration between at least two classes of mutations: class I, exemplified by constitutively activated tyrosine kinases or their downstream effectors, which confer a proliferative and/or survival advantage to hematopoietic cells; and class II mutations, that result in loss of function of transcription factors that are important for normal hematopoietic differentiation. The activating V617F mutation of JAK2 tyrosine kinase has been recently described as a common event in MPD. This mutation seems to disrupt the JH2 inhibitory regulation of JAK2, leaving the enzyme constitutively active. The same mutation was also found in a small number of patients with either atypical MPD or MDS. However, its incidence in AML is unclear. We investigated the incidence of the tyrosine kinase JAK2 mutation V617F in patients with AML, and its association with other factors with a prognosis meaning in this disease. We focused in cases with normal karyotype, which account for more than 40% of all AML, and that represent a therapeutic challenge because of lack of established therapy protocols. Material and methods. We analyzed 7 cell lines and 218 samples of AML patients at diagnosis: 38 included in the favorable prognosis group, 143 in the intermediate (137 with normal karyotype), and 45 in the unfavorable. DNA was extracted by standard procedures from bone marrow. V617F genotyping was performed by amplification refractory mutation system (ARMS) as previously described (Jones el al., 2005). Mutations of FLT3 (ITD and D835) were also analyzed. Results. We found the mutation V617F in 4.5% of overall AML cases (10/225), and in 4.4% of AML patients with normal karyotype (6/137). A high prevalence was observed in patients with the M2 subtype (7/61, 11.5%): 5 cases had a normal karyotype, one t(8;21), and another inv(16). Taking into account cases with AML-M2 and normal karyotype, the incidence of the V617F mutation was 19% (5/26). The mutation was heterozygous in most cases; we only detected homozygous mutation in 2 cell lines with AML-M6 and M7, both presenting a complex karyotype and 3q rearrangements. One of the cell lines had an addition on 9p24, supporting previous data that imply that 9pLOH, caused by mitotic recombination, could be responsible for the transition from heterozygous to homozygous V617F. Interestingly, one of the cases with M2 had both FLT3 and JAK2 mutations. FLT3 mutations were no detected in the other cases. Genetic evidence in vivo and studies on proliferation and phosphorylation in vitro indicate that the V617F mutation confers a proliferative and survival advantage. Patients carrying the V617F mutation had significantly longer disease duration and more complications in MPD, suggesting a possible prognosis meaning in AML patients. In conclusion, we have identified an acquired JAK2 mutation, which induces a constitutive JAK2 signaling, in 19% of patients with AML-M2 and a normal karyotype. Our data suggest that this activating JAK2 mutation may play an important role in the leukemogenesis mechanism of AML, and could contribute to define a subgroup within patients with AML and normal karyotype. Nevertheless, the presence of the V617F mutation opens the possibility for new targeted therapies in AML. Further studies of the clinical significance of the JAK2 mutation in AML-M2 will be of interest.

Patients with 3q21q26 rearrangements seem to share similar clinicopathologic features and a common molecular mechanism, leading to MDS or AML. The overexpression of EVI1 (3q26) has been implicated in the dysplasia that characterizes this subset of myeloid neoplasias with poor prognosis. However, EVI1 overexpression had been detected in 9% of AML cases without 3q26 abnormalities; and several 3q21q26 cases with no EVI1 overexpression had been reported. Our aim was to characterize different subgroups in 3q rearrangements, and to define the gene expression profile of these groups, specially in the group with 3q21q26 rearrangements. We analyzed 44 samples (6 cell lines and 38 samples of patients) with 3q rearrangements at G-band level, 36 had AML and 8 MDS. As a control, we analyzed 6 samples of patients with 3q rearrangements and lymphoid neoplasias, a cDNA panel of normal human tissues (Clontech), and 10 normal bone marrow samples. FISH analysis was performed using six BACs: PR11-390G14, RP11-475N22 (GATA2), RP11-689D3 (RPN1) on 3q21; RP11-82C9 (EVI1), RP11-115B16 (MDS1) and RP11-196F13 (TRAIL) on 3q26; and a centromere probe. Expression of GATA2, MDS1-EVI1, and EVI1 was measured by a quantitative real-time RT-PCR (Taqman) assay (Applied Biosystems). Samples were first classified attending at the G-banding. FISH analysis showed the heterogeneity of breakpoints in these cases, and the samples were grouped as 3q21 (5 cases), 3q26 (11 cases), 3q21q26 (16 cases) and other 3q rearrangements (12 cases). Real time RT-PCR showed overexpression (OE) of EVI1 and/or GATA2 in 75% of all the cases (33/44), suggesting the important role of these genes in patients with 3q rearrangements and myeloid neoplasias. In patients with lymphoid neoplasias and 3q rearrangements no genes were found to be OE. A different expression profile was found in the groups analyzed (Table 1). In the 3q21 group, 40% of the patients had GATA2 OE, but no EVI1 or MDS1-EVI1 OE was found. Our data show that EVI1 OE could be specific of the groups 3q26 (64%) and 3q21q26 (69%). The number of cases with EVI1 OE was significantly higher in patients with 3q rearrangements than in those with complex karyotype (data not show). In the group with 3q21q26, EVI1 and/or GATA2 were overexpressed in 87.5% of the cases (14/16). In conclusion, we have detected a different expression profile of the genes analyzed among samples with different 3q rearrangements. Our results suggest a more complex mechanism in patients with 3q21q26 rearrangements, in which GATA2 overexpression may contribute with EVI1 to the leukemogenic progress. Further molecular studies are in progress to analyze the different EVI1 transcripts.

The presence of recurrent cytogenetic aberrations in AML has diagnostic and prognostic value; however, more than 40% of patients have a normal karyotype. Although the presence of FLT3 mutations, or overexpression (OE) of WT1, EVI1 or BAALC, have been reported to discriminate in this group those with a worse prognosis, more molecular markers are necessary to achieve treatment stratification in this heterogeneous group. The mice homolog of EVI1 (3q26) has been reported to have a role in the HSC proliferation through Gata2 expression, a downstream target gene of the Evi1 oncogene. Therefore, GATA2 (3q21), a transcription factor with a relevant role in hematopoiesis, could be a candidate gene in the leukemogenic transformation of AML. Our aim was to determinate the incidence of GATA2 OE in AML, and its association with other well-defined prognostic factors. We analyzed 7 cell lines and 192 samples of AML patients at diagnosis: 61 were in the favorable prognosis group, 82 in the intermediate (68 with normal karyotype) and 56 in the unfavorable (27 with 3q rearrangements). Expression of GATA2, MDS1/EVI1, EVI1 and WT1 were measured by a RTQ-PCR Taqman assay (Applied Biosystems). Mutations of FLT3 (ITD and D835) were also analyzed. Gene expression arrays were performed with the Human 19K Oligo Array (Center for Applied Genomics, University of New Jersey). We detected a different expression profile among the 3 prognosis groups: 42% and 12% of all AML cases had GATA2 and EVI1 OE respectively, and the frequency increased in the poorer prognosis groups (Table 1). GATA2 OE was detected in 46% of cases with normal karyotype, and was more frequent among samples with FLT3-ITD (68% vs 34%; p=0.0021). Among AML-M1 cases, the number of samples with FLT3-ITD and GATA2 OE was twice than in samples without FLT3 mutations (78% vs 39%). All patients with AML-M1, FLT3-ITD and GATA2 OE had WT1 OE. Gene expression array technology was performed to compare 3 groups of AML-M1: I) FLT3 no mutated and GATA2 OE; II) FLT3 no mutated and GATA2 no OE; and III) FLT3-ITD and GATA2 OE. Unsupervised analysis clearly classified these 3 groups, suggesting they could be prognostically relevant AML subgroups, and allowing the detection of 38 candidate genes that were differentially expressed. Our results show that GATA2 OE is a common event in AML cases with normal karyotype (46%). The mutational pattern FLT3-ITD/GATA2/WT1 could define a subgroup of patients with normal karyotype and AML-M1, with a different expression pattern.