Randomised Introduction of 2-CDA as Intensification during Consolidation for Children with High-risk AML--results from Study AML-BFM 2004.

BACKGROUND: The outcome in children and adolescents with high-risk (HR) acute myeloid leukemia (AML) is still unsatisfactory. Therefore, in study AML-BFM 2004 we aimed to improve outcome of HR-patients by adding moderately dosed 2-Chloro-2-Deoxyadenosine (2-CDA) to the respective consolidation treatment backbone without increasing toxicity. The aim was to improve prognosis especially in FAB M4/M5/MLL patients, who represent the largest subgroup of HR patients. PATIENTS AND METHODS: In total, 343 children and adolescents with HR-AML were randomized to receive or not 2-CDA (6 mg/m²/d, days 1, 3) in combination with cytarabine/idarubicine (AI=500 mg/m² cytarabine 5 days continuous infusion plus 7 mg/m²/d idarubicin, days 3 and 5). RESULTS: RESULTS for patients of the AI/2-CDA arm (n=168) vs. the AI-arm (n=175) were similar: 5-year overall survival 68±4 vs. 72±4%, plogrank=0.38, event-free survival 53±4 vs. 49±4%, plogrank=0.77; cumulative incidence of relapse at 5 years: 35±4 vs. 37±4%, p(Gray)=0.89. RESULTS in patients with MLL rearrangement or FAB M4/M5 were also similar in the treatment groups. In addition, toxicities did not differ between the two arms. CONCLUSION: We conclude that additional, moderate dose 2-CDA does not improve prognosis in HR-patients when given during consolidation treatment. Its effect might be too low in this multidrug regimen, where the strongest effects are achieved during induction, or the chosen dose of 2-CDA might have been too low.

Development of a curative treatment within the AML-BFM studies.

The first multicenter treatment study for AML in childhood in Germany was performed from 1978 onwards. The therapy plan was designed similar to that for the acute lymphoblastic leukaemia (ALL). The drugs with the highest efficacy in AML, cytarabine cutting catara-bine and anthracyclines, were combined during induction and consolidation, followed by preventive cranial irradiation and maintenance therapy similar to that in ALL. The remission rate of the initial study was 80%, and the 5-year survival rate increased from less than 10% before 1970 to 40%. 5 subsequent trials have further increased the 5-year survival to now 70% and even 90% in the subgroup of core-binding factor leukaemias by using an intensified and optimised treatment schedule.The AML-BFM studies were the only prospective study sequence testing the benefit of cranial irradiation. Results from study -87 including the non-randomized patients showed an increased risk of CNS and/or bone marrow relapses in non-irradiated patients. Later on there was evidence that 12 Gy resulted in the same relapse rate as 18 Gy. The AML-BFM studies always used the experience from the previous study to optimize the next study. This approach was essential together with improved supportive treatment and experience of the medical staff for the step-wise and considerable increase of longterm survival within the 6 subsequent AML-BFM studies.

Genotype-outcome correlations in pediatric AML: the impact of a monosomal karyotype in trial AML-BFM 2004.

We conducted a cytogenetic analysis of 642 children with de novo acute myeloid leukemia (AML) treated on the AML-Berlin-Frankfurt-Münster (BFM) 04 protocol to determine the prognostic value of specific chromosomal aberrations including monosomal (MK+), complex (CK+) and hypodiploid (HK+) karyotypes, individually and in combination. Multivariate regression analysis identified in particular MK+ (n=22) as a new independent risk factor for poor event-free survival (EFS 23±9% vs 53±2% for all other patients, P=0.0003), even after exclusion of four patients with monosomy 7 (EFS 28±11%, P=0.0081). CK+ patients without MK had a better prognosis (n=47, EFS 47±8%, P=0.46) than those with MK+ (n=12, EFS 25±13%, P=0.024). HK+ (n=37, EFS 44±8% for total cohort, P=0.3) influenced outcome only when t(8;21) patients were excluded (remaining n=16, EFS 9±8%, P<0.0001). An extremely poor outcome was observed for MK+/HK+ patients (n=10, EFS 10±10%, P<0.0001). Finally, isolated trisomy 8 was also associated with low EFS (n=16, EFS 25±11%, P=0.0091). In conclusion, monosomal karyotype is a strong and independent predictor for high-risk pediatric AML. In addition, isolated trisomy 8 and hypodiploidy without t(8;21) coincide with dismal outcome. These results have important implications for risk stratification and should be further validated in independent pediatric cohorts.

AML patients with Down syndrome have a high cure rate with AML-BFM therapy with reduced dose intensity.

Despite improved prognosis in acute myelogenous leukaemia (AML) children with Down syndrome (DS), therapy-related toxicity remained a problem. We compared 67 DS patients from study AML-BFM 98 with 51 DS patients of the previous study AML-BFM 93, and the non-DS groups of both studies. Compared to non-DS patients, DS patients were treated with reduced anthracycline doses, without high-dose cytarabine/mitoxantrone and without cranial irradiation. AML-DS patients were in median 1.8 years old, and 102/118 (86%) showed the typical morphology of acute megakaryoblastic leukaemia. In study 93, seven DS patients did not receive AML-specific chemotherapy, and treatment modifications were more common. Results improved significantly for patients treated in study 98 with a 3-year survival of 91+/-4 vs 70+/-7% in study 93 (P=0.001). There were no differences in outcome concerning the age groups 0-

Drug-induced immunophenotypic modulation in childhood ALL: implications for minimal residual disease detection.

Assessment of minimal residual disease (MRD) by flow cytometry is considered to be based on the reproducibility of the leukemic immunophenotype detected at diagnosis. However, we previously noticed modulation of surface antigen expression in acute lymphoblastic leukemia (ALL) during the early treatment. Hence, we investigated this in 30 children with B-cell precursor ALL consecutively enrolled in the AIEOP-BFM ALL 2000 protocol. Quantitative expression of seven antigens useful in MRD monitoring was studied at diagnosis and compared to that measured at different time points of remission induction therapy. Downmodulation in the expression of CD10 and CD34 occurred at follow-up. By contrast, upmodulation of CD19, CD20, CD45RA, and CD11a was observed, while the expression of CD58 remained stable. Despite this, we could unambiguously discriminate leukemic cells from normal residual B cells. This holds true when bone marrow (BM) samples from similarly treated T-ALL patients, but not from healthy donors, were used as reference. Our results indicate that immunophenotypic modulation occurs in ALL during the early phases of BFM-type protocols. However, the accuracy of MRD detection by flow cytometry seems not negatively affected if adequate analysis protocols are employed. Investigators should take this phenomenon into account in order to avoid pitfalls in flow cytometric MRD studies.

Overexpression of the pseudoautosomal gene MIC2 in Ewing's sarcoma and peripheral primitive neuroectodermal tumor.

Ewing's Sarcoma (ES), the second most frequent bone tumor in childhood and adolescence, and the probably closely related peripheral primitive neuroectodermal tumor (pPNET) share a unique cytogenetic translocation between chromosomes 11 and 22. Both of them expose high amounts of a glycoprotein on their cell surface, which can be specifically detected by the mAb HBA-71. The cDNA coding for the HBA-71 antigen was isolated by screening a cDNA expression library constructed from a pPNET-derived cell line. Nucleotide sequencing revealed the HBA-71 antigen to be the product of the pseudoautosomal gene MIC2 previously identified by the mAb 12E7 in haematopoietic cells. This antigen is a glycoprotein with a molecular weight of about 29,000 and is expressed in low amounts in most human cell lines and probably normal tissues and tumors with only a few exceptions. In T-cells the antigen is involved in cell adhesion processes. In ES- and pPNET-derived cell lines MIC2 expression is significantly enhanced. No gross changes in posttranslational modification could be observed. The high expression results in easy and specific detection of the antigen in immunocytochemical analysis of paraffin embedded tissue sections making HBA-71 a useful tool in tumor diagnosis.

PAX5/ETV6 fusion defines cytogenetic entity dic(9;12)(p13;p13).

Recurrent chromosomal abnormalities present in malignant cells often define subentities with unique biological and clinical features. The molecular identification of genes involved in genetic alterations has led to the characterization of fusion genes with neoplastic properties. However, for many nonrandom translocations including the dic(9;12)(p11-13;p11-12), the molecular equivalent has not as yet been identified. The dicentric translocation dic(9;12) is a recurrent chromosome abnormality that accounts for close to 1% of childhood acute lymphoblastic leukemia (ALL). This specific alteration occurs almost exclusively in B-progenitor ALL, and unlike many other nonrandom translocations, is associated with an excellent prognosis. In this work, we provide strong evidence that the PAX5/ETV6 fusion transcript defines the clinical and biological entity that is associated with the presence of a dic(9;12) chromosome. As the PAX5 and ETV6 genes are localized at 9p13 and 12p13, respectively, the cytogenetic description of the dic(9;12)-PAX5/ETV6 rearrangement should be refined to dic(9;12)(p13;p13).

Incidence and relevance of secondary chromosome abnormalities in childhood TEL/AML1+ acute lymphoblastic leukemia: an interphase FISH analysis.

The aim of the present study was to determine the frequency and clinical relevance of the most common secondary karyotype abnormalities in TEL/AML1+ B-cell precursor acute lymphoblastic leukemia (ALL) as assessed with fluorescence in situ hybridization (FISH) analyses. Screening of 372 patients who were enrolled in two consecutive Austrian childhood ALL multicenter trials identified 94 (25%) TEL/AML1+ cases. TEL deletions, trisomy 21 and an additional der(21)t(12;21) were detected in 52 (55%), 13 (14%) and 14 (15%) TEL/AML1+ patients, respectively. The 12p aberrations (P=0.001) and near tetraploidy (P=0.045) were more common in TEL/AML1+ patients, whereas the incidence of diploidy, pseudodiploidy, hypodiploidy, low hyperdiploidy, near triploidy, del(6q), chromosome 9 and 11q23 abnormalities was similar among TEL/AML1+ and TEL/AML1- patients. None of the TEL/AML1+ patients had a high hyperdiploid karyotype. Univariate analysis indicated that among TEL/AML1+ patients those with a deletion of the nontranslocated TEL allele had a worse prognosis than those without this abnormality (P=0.034). We concluded that the type and incidence of the most common secondary aberrations in TEL/AML1+ ALL can be conveniently identified with little additional effort during interphase screening with appropriate TEL and AML1 FISH probes. We also provided preliminary evidence that the deletion of the nontranslocated TEL allele may adversely influence the clinical course of TEL/AML1+ ALL.

CD99 expression in T-lineage ALL: implications for flow cytometric detection of minimal residual disease.

Expression of CD99 is higher on immature than on mature T cells. We postulated that this marker could be used to assess minimal residual disease (MRD) in T-lineage acute lymphoblastic leukemia (T-ALL). In diagnostic bone marrow (BM) samples from 27 children with T-ALL, expression of CD99 on leukemic lymphoblasts by flow cytometry was in median 7.7 times higher than on normal T lymphocytes from within the same sample. In 85% of cases, leukemic MFI values were higher than the mean MFI+2 s.d. of normal populations. We applied CD99 to study MRD in 39 follow-up samples from 15 consecutive T-ALL patients, and compared the results with those obtained with the well-established MRD-marker terminal deoxynucleotidyl transferase (TdT). Either antibody was combined in four-color flow cytometry with CD7, surfaceCD3, and cytoplasmicCD3. We found that CD99 was a valid complement to TdT in quantifying T-ALL MRD. Given a considerable interpatient variability, CD99 could be favorably used in nine patients, and TdT in other five patients. Both approaches showed a similar very low nonspecific background throughout 12 weeks from diagnosis (in median 0.002% of nucleated BM cells in patients with non-T ALL). We conclude that CD99 is a highly informative tool for MRD detection in T-ALL, bearing the advantage of surface expression in contrast to TdT.

Multiparameter phenotype mapping of normal and post-chemotherapy B lymphopoiesis in pediatric bone marrow.

We studied the differentiation profiles of B cell precursors (BCP) in normal and post-chemotherapy pediatric bone marrow (BM) using multiparameter flow cytometry. The goal of our study was to draw a comprehensive phenotypic map of the three major maturational BCP stages in BM. By correlating lineage-associated markers, CD45RA, and several adhesion molecules, the stage-specific patterns were found to differ in certain details from previously published concepts. Among the earliest BCP, a subset of CD34+ CD10(lo) precursors was repeatedly observed in addition to the well characterized CD34+ CD10(hi) CD19+ majority of cells. Only two-thirds of these CD34+ CD10(lo) cells expressed CD19. However, uniformity of phenotypic features, absence of T lineage markers, and the regeneration kinetics after chemotherapy suggest the B lineage affiliation of the CD34+ CD10(lo) precursors in general. In the more mature BCP, expression of CD10, CD20, cytoplasmic and surface mu chains (c mu and s mu) was observed to overlap more than previously recognized. We found that CD20 and c mu appear early during B cell ontogeny (already on CD34+ BCP), and that CD10 is lost late, following the onset of s mu expression. Differences between normal and post-chemotherapy BM specimens regarding the phenotypic appearance of BCP were exclusively due to differences in the subset composition, as post-chemotherapy samples showed a preponderance of immature stages. Our observations may build a framework for comparing leukemic cells with their normal counterparts to define possible leukemia-associated aberrations useful for residual disease studies.

Detection of residual disease in pediatric B-cell precursor acute lymphoblastic leukemia by comparative phenotype mapping: a study of five cases controlled by genetic methods.

We recently investigated samples of pediatric B-cell precursor acute lymphoblastic leukemia (BCP-ALL) and normal bone marrow (BM). We found that leukemic blasts, compared to their physiologic counterpart cells, frequently display aberrant phenotypes with respect to levels of expression of certain antigens. Using multiparameter flow cytometry, these differences enabled us to trace leukemic cells admixed to normal BM, which suggested that this approach might be a useful strategy for minimal residual disease detection. In the present study, we used the same multiparameter approach ("comparative phenotype mapping") to prove that such quantitative phenotypic differences really exist between malignant and normal BCP when simultaneously present in the BM. We demonstrate this by five exemplary follow-up BM samples from patients with BCP-ALL, all of which showed phenotypically aberrant cells according to levels of expression of CD10, CD11a, CD19, CD34, CD44, or CD45RA, as well as according to altered orthogonal light scattering properties. We confirmed the leukemic nature of these cells by polymerase chain reaction-based detection of bcr1/abl transcripts, and of leukemia clone-specific immunoglobulin heavy chain rearrangements in only the suspicious cells when sorted by flow cytometry, but not in normal BCP or non-B cells. Comparative phenotype mapping thus allows one to distinguish between normal and leukemic cells, and we show that it may enable rapid, specific, and quantitative detection of residual/resurgent leukemia in BCP-ALL.

Comparative phenotype mapping of normal vs. malignant pediatric B-lymphopoiesis unveils leukemia-associated aberrations.

Leukemic cells of B-lineage acute lymphoblastic leukemia (ALL) are regarded as the malignant counterparts of immature, physiologic B cell precursors (BCPs). To determine whether phenotypic differences exist between these corresponding cell types, we investigated samples of normal pediatric bone marrow (n=30) as well as of B-precursor ALL at diagnosis (n=53; common and pre-B subtype). Using three-color multiparameter flow cytometric analysis, we compared the leukemic populations with the physiologic BCPs of corresponding maturity with respect to the intensity with which they expressed a series of antigens. In some of these antigens, leukemia-associated aberrations were frequently observed. In particular, overexpression of CD10 was displayed by 65% of ALL samples, whereas 58% of leukemic cases aberrantly exhibited very low or no CD45RA expression. Regarding CD11a and CD44, 47% and 35% of ALL populations were aberrant as defined by either the absence or significant overexpression of the antigen. In contrast, antigen densities of CD49d, CD49e, and CD99 on leukemic cells were in the normal range of values for BCPs. Combining the patterns of frequently aberrant markers in a comprehensive analysis, we were able to identify individual phenotypic leukemic cell aberrations in up to 98% of investigated cases. CD10 and/or CD45RA were aberrant in 86% of cases overall, emphasizing the high discriminative potential of these two markers. Using comparative phenotype mapping based on quantitatively aberrant, leukemia-associated antigenic patterns, we were able to detect leukemic blasts among normal bone marrow cells at frequencies as low as 10(-5). We speculate that our approach may have a profound impact on the development of new strategies for minimal residual disease investigations in patients with BCP-ALL.

Detection of residual disease in pediatric B-cell precursor acute lymphoblastic leukemia by comparative phenotype mapping: method and significance.

The present review summarizes our efforts in developing a novel immunologic approach ("Comparative Phenotype Mapping") targeted at assessing minimal residual disease (MRD) in B-cell precursor (BCP) acute lymphoblastic leukemia (ALL) patients. The method relies on quantitatively aberrant, leukemia-associated antigen expression patterns which allow to discriminate leukemic from normal BCP using a limited panel of antibody combinations and multidimensional flow cytometry. In an analysis of 63 follow up bone marrow samples of patients with BCP-ALL we show that this approach enables to efficiently detect MRD. Further clinical observation revealed that the patients which were MRD-positive by flow cytometry (although in morphological remission) had a very high probability of early disease recurrence compared to the good chances of a relapse-free survival (RFS) in the MRD-negative cohort (RFS 0.0 vs. 0.76 at 3 years). Comparative Phenotype Mapping thus proves to be a reliable method for MRD detection in BCP-ALL. Concluding remarks relate to the optional applications of the method as well as to future perspectives. An ongoing large prospective study which we are now conducting on the basis of Comparative Phenotype Mapping will clarify the clinical significance of MRD detection in ALL patients by this method, and will determine its value compared to related as well as molecular-genetic techniques.

CD2-positive B-cell precursor acute lymphoblastic leukemia with an early switch to the monocytic lineage.

Switches from the lymphoid to myeloid lineage during B-cell precursor acute lymphoblastic leukemia (BCP-ALL) treatment are considered rare and thus far have been detected in MLL-rearranged leukemia. Here, we describe a novel BCP-ALL subset, switching BCP-ALL or swALL, which demonstrated monocytosis early during treatment. Despite their monocytic phenotype, 'monocytoids' share immunoreceptor gene rearrangements with leukemic B lymphoblasts. All swALLs demonstrated BCP-ALL with CD2 positivity and no MLL alterations, and the proportion of swALLs cases among BCP-ALLs was unexpectedly high (4%). The upregulation of CEBPα and demethylation of the CEBPA gene were significant in blasts at diagnosis, prior to the time when most of the switching occurs. Intermediate stages between CD14(neg)CD19(pos)CD34(pos) B lymphoblasts and CD14(pos)CD19(neg)CD34(neg) 'monocytoids' were detected, and changes in the expression of PAX5, PU1, M-CSFR, GM-CSFR and other genes accompanied the switch. Alterations in the Ikaros and ERG genes were more frequent in swALL patients; however, both were altered in only a minority of swALLs. Moreover, switching could be recapitulated in vitro and in mouse xenografts. Although children with swALL respond slowly to initial therapy, risk-based ALL therapy appears the treatment of choice for swALL. SwALL shows that transdifferentiating into monocytic lineage is specifically associated with CEBPα changes and CD2 expression.

T-cell receptor Vß skewing frequently occurs in refractory cytopenia of childhood and is associated with an expansion of effector cytotoxic T cells: a prospective study by EWOG-MDS.

Immunosuppressive therapy (IST), consisting of antithymocyte globulin and cyclosporine A, is effective in refractory cytopenia of childhood (RCC), suggesting that, similar to low-grade myelodysplastic syndromes in adult patients, T lymphocytes are involved in suppressing hematopoiesis in a subset of RCC patients. However, the potential role of a T-cell-mediated pathophysiology in RCC remains poorly explored. In a cohort of 92 RCC patients, we prospectively assessed the frequency of T-cell receptor (TCR) ß-chain variable (Vß) domain skewing in bone marrow and peripheral blood by heteroduplex PCR, and analyzed T-cell subsets in peripheral blood by flow cytometry. TCRVß skewing was present in 40% of RCC patients. TCRVß skewing did not correlate with bone marrow cellularity, karyotype, transfusion history, HLA-DR15 or the presence of a PNH clone. In 28 patients treated with IST, TCRVß skewing was not clearly related with treatment response. However, TCRVß skewing did correlate with a disturbed CD4(+)/CD8(+) T-cell ratio, a reduction in naive CD8(+) T cells, an expansion of effector CD8(+) T cells and an increase in activated CD8(+) T cells (defined as HLA-DR(+), CD57(+) or CD56(+)). These data suggest that T lymphocytes contribute to RCC pathogenesis in a proportion of patients, and provide a rationale for treatment with IST in selected patients with RCC.

CD99 ligation upregulates HSP70 on acute lymphoblastic leukemia cells and concomitantly increases NK cytotoxicity.

CD99 is present in many human cell types, including high-level surface expression on pediatric B and T leukemias and Ewing tumors (ETs). On B lymphocytes and respective malignancies, its level decreases with the stage of maturation. Inter-individual variability of CD99 on B-cell precursor acute lymphoblastic leukemia (BCP-ALL) blasts was shown recently to be associated with distinct cytogenetic backgrounds. However, CD99 targets remain mainly unknown. Here, we show that administration of an anti-CD99 antibody to B- and T-leukemia cell lines induces heat shock protein 70 (HSP70), both on the cell surface and in the cytoplasm. Investigation of primary BCP-ALL cells rendered similar results. Intriguingly, CD99-induced modulation of HSP70 on ET cells had profiles different from that on leukemia cells. Since HSP70 expression on tumor cells is a prerequisite for natural killer (NK) cell-mediated tumor lysis, we hypothesized that CD99-induced HSP70 may allow targeting of some CD99-positive malignancies via NK-cell cytotoxicity. Our experiments with NK92 cell line demonstrated that leukemia cells with upregulated HSP70 can be successfully killed by effector cells. We consider our data as a new view of CD99 functions and as a basis for the development of a potential anti-tumor strategy based on heat-shock protein activation via CD99 triggering.