Flow cytometric analysis of de novo acute lymphoblastic leukemia in childhood: report from the Japanese Pediatric Leukemia/Lymphoma Study Group

Although the antigen expression patterns of childhood acute lymphoblastic leukemia (ALL) are well known, little attention has been given to standardizing the diagnostic and classification criteria. We retrospectively analyzed the flow cytometric data from a large study of antigen expression in 1,774 children with newly diagnosed ALL in JPLSG. T- and B-lineage ALL accounted for 13 and 87% of childhood ALL cases, respectively. Cytoplasmic CD3 and CD7 antigens were positive in all T-ALL cases. More than 80% of T-ALL cases expressed CD2, CD5 and TdT. In B-lineage ALL, the frequencies of early pre-B, pre-B, transitional pre-B and B-ALL were 81, 15.5, 0.6 and 2.9%, respectively. More than 90% of early pre-B ALL cases expressed CD19, CD79a, CD22, CD10 and TdT. CD34 was expressed in three-fourths of early pre-B ALL cases. The frequencies of TdT and CD34 expression were lower in pre-B ALL than in early pre-B ALL. B-ALL showed less frequent expression of CD22, CD10, CD34 and TdT than other B-lineage ALL cases. Expression of CD13 and CD33, aberrant myeloid antigens, was significantly more frequently associated with B-lineage ALL than with T-ALL. Based on this retrospective study of antigen expression in 1,774 de novo childhood ALL cases in JPLSG, we propose standardized clinical guidelines for the immunophenotypic criteria for diagnosis and classification of pediatric ALL.     

Expression of adhesion molecules in childhood B-lineage-cell neoplasms

We analyzed the expression pattern of adhesion molecules including beta 1-integrins (CD49c, CD49d, CD49e, CD49f), beta 2-integrins (CD11a, CD11b, CD11c), CD44, and CD54 in 141 children with B-cell precursor acute lymphoblastic leukemia (pre-B ALL) and in 21 children with B-cell ALL/non-Hodgkin's lymphoma (B-ALL/NHL). The frequencies of CD11a, CD49f, and CD44 expression were significantly higher in CD34+ pre-B ALL than in CD34- pre-B ALL. Although CD49d, CD49e, and CD44 were less frequently expressed in B-ALL/NHL than in pre-B ALL, the expression of CD11a and CD54 were more frequent in B-ALL/NHL. In pre-B ALL, expression of CD11a positively correlated with that of CD11b (P < .05) and CD54 (P < .01), and CD49c positively correlated with CD49f (P < .01). Of the clinical parameters of patients with pre-B ALL, expression of CD11a was associated with a low leukocyte count (P < .05). The presence of CD54 on the cell surface was an independent factor indicating a poor prognosis. The estimated 5-year event-free survival was 42.3% for CD54+ (n = 31) compared with 70.3% for CD54- patients (n = 38) (P < .05). These findings demonstrated that expression of adhesion molecules is dependent on the phenotype of B-lineage cells and that the expression of some of these molecules has clinical significance.     

Pre-B acute lymphoblastic leukemia cells may induce T-cell anergy to alloantigen

Even if neoplastic cells express tumor associated antigens they still may fail to function as antigen presenting cells (APC) if they lack expression of one or more molecules critical for the induction of productive immunity. These cellular defects can be repaired by physiologic activation, transfection, or fusion of tumor cells with professional APC. Although such defects can be repaired, antitumor specific T cells may still fail to respond in vivo if they may have been tolerized. Here, human pre-B cell acute lymphoblastic leukemia (pre-B ALL) was used as a model to determine if primary human tumor cells can function as alloantigen presenting cells (alloAPC) or alternatively whether they induce anergy. In the present report, we show that pre-B cell ALL express alloantigen and adhesion molecules but uniformly lack B7-1 (CD80) and only a subset express B7-2 (CD86). Pre-B ALL cells are inefficient or ineffective alloAPC and those cases that lack expression of B7-1 and B7-2 also induce alloantigen specific T- cell unresponsiveness. Under these circumstances, T-cell unresponsiveness could be prevented by physiologic activation of tumor cells via CD40, cross-linking CD28, or signaling through the common gamma chain of the interleukin-2 receptor on T cells. Taken together, these results suggest that pre-B ALL may be incapable of inducing clinically significant T-cell-mediated antileukemia responses. This defect may be not only due to their inability to function as APC, but also due to their potential to induce tolerance. Attempts to induce clinically significant antitumor immune responses may then require not only mechanisms to repair the antigen presenting capacity of the tumor cells, but also reversal of tolerance.     

Prethymic phenotype and genotype of pre-T (CD7+/ER-)-cell leukemia and its clinical significance within adult acute lymphoblastic leukemia

Pretreatment blast cells from 739 adults with acute lymphoblastic leukemia (ALL) were immunophenotyped as part of a prospective treatment protocol study. Among 192 patients (26%) with T lineage ALL, 47 (6%; 24% of T lineage ALL) had lymphoblasts without sheep erythrocyte rosette formation, but with pan-T antigen CD7 on the membrane and intracellular CD3 proteins mostly in perinuclear accumulation. The T-cell surface antigens CD5 and/or CD2 and focal acid phosphatase were additional markers of this subgroup traditionally called pre-T ALL, whereas thymocyte antigen CD1 as well as CD4 and CD8 antigens were not expressed. Hematopoietic progenitor cell markers, namely terminal deoxynucleotidyl transferase (TdT), and in part common ALL antigen (CD10), HLA-DR antigens, and/or My-10 (CD34), a unique antigen of marrow cells absent in thymus cells, further characterized this immature T-ALL form of putative prothymocytic phenotype (CD7+/intracellular CD3+/TdT+/My-10+/-/HLA-DR+/-/CD10+/-). The prethymic T cell character was supported by germ-line T-cell receptor beta genes found in 21 of 36 patients analyzed. In five cases only T gamma-chain genes were rearranged. Fifteen patients, however, had rearrangements of both T beta and T gamma genes. Immunoglobulin heavy chain genes were rearranged only in two cases. Pre-T ALL differed significantly from E-rosette+ T-ALL in some presenting clinical features, namely mediastinal mass, lymphoadenopathy, and platelet count, and independently of clinical factors in prognosis (P = .02, median remission duration: 15.7 v 33.5 months, and P = .02, median survival time: 24.6 v 50.7 months). We conclude that ALL classification based solely on T- or B-cell lineage affiliation is not sufficient but needs further subdivision according to relevant maturation stages as exemplified here within the T-cell axis. The putative prethymic T cell progenitor phenotype described might help elucidate the sequence of genetic events that commit normal hematopoietic cells to the T-cell lineage.     

Lack of CD45 antigen on blast cells in childhood acute lymphoblastic leukemia is associated with chromosomal hyperdiploidy and other favorable prognostic features.

The leukocyte common antigen (CD45) was detected on the surface of leukemic cells in 217 (87%) of 249 cases of newly diagnosed childhood acute lymphoblastic leukemia (ALL). All 55 cases of T-lineage ALL, compared with 159 of 191 B-lineage cases, expressed the CD45 antigen (P = .0005). The frequency of CD45 expression did not differ between cases of early pre-B (CD19+, cytoplasmic mu-) and pre-B (CD19+, cytoplasmic mu+) ALL. Cases of ALL lacking CD45 had significantly lower leukocyte counts (P = .002) and serum lactic dehydrogenase (LDH) levels (P = .007) and were more likely to have leukemic cell hyperdiploidy greater than 50 (P less than .0001) or a DNA index greater than 1.15 (P less than .0001), as compared with cases positive for the antigen. Of the 130 patients whose follow-up duration was sufficient for analysis of event-free survival, the 53 with the highest levels of CD45 expression (greater than or equal to 90%) were the most likely to have an adverse event on intensive multiagent chemotherapy. Patients without detectable CD45 had a negligible risk of failure. This study suggests a relationship between the expression of the CD45 antigen on leukemic lymphoblasts and other biologic factors that influence prognosis in ALL.     

Clinical and biologic features of childhood T-cell leukemia with the t(11;14)

Cytogenetic analysis of cells from 622 consecutive patients with newly diagnosed acute lymphoblastic leukemia (ALL) and successful G-banding chromosome studies disclosed seven cases with the t(11;14)(p13;q11) and one with the t(11;14)(p15;q11). Leukemia cells in all eight cases had a T-cell immunophenotype. The t(11;14)(p13;q11) occurred in 6.8% and the t(11;14)(p15;q11) in 1% of T-cell ALL cases (n = 103). The t(11;14) was associated with presenting clinical features typical of T-cell ALL: male predominance (n = 6), age greater than 10 years (n = 3), hyperleukocytosis (white blood cells greater than 100 x 10(9)/L, n = 5), relatively high hemoglobin level (median, 10.8 g/dL), high serum lactic dehydrogenase level (median, 3248 U/L), presence of mediastinal mass (n = 6), and central nervous system leukemia (n = 2). While there were no significant differences in presenting features between T-cell ALL cases with or without the t(11;14), leukemic cells from patients with the translocations were more likely to coexpress CD4 and CD8 antigens (6 of 6 v 35 of 86 cases tested, P less than .05). Adverse events have occurred in six patients: three central nervous system relapses [including the one with t(11;14)(p15;q11)], two secondary acute myeloid leukemia, and one hematologic relapse. Our results indicate that the t(11;14)(p13;q11) occurs exclusively in T-cell malignancies of intermediate- or late-stage thymocyte differentiation. Additional studies are needed to determine the prognostic implications of these translocations.     

Heterogeneity of presenting features and their relation to treatment outcome in 120 children with T-cell acute lymphoblastic leukemia

Presenting features of 120 consecutive children with T-cell acute lymphoblastic leukemia (ALL), representing 15% of all patients diagnosed as having ALL during the study period, were analyzed to determine relationships with treatment outcome. Patients' ages ranged from 1.7 to 18.8 years (median, 10.3 years) and their leukocyte counts from 1.7 to 1,070 x 10(9)/L (median, 100 x 10(9)/L). Central nervous system (CNS) leukemia was present in 12.5% of the cases, a mediastinal mass in 61%, and L2 lymphoblast morphology in 32%. A relatively high proportion of cases, 26%, had normal karyotypes at presentation. Of the cases tested, membrane CD1 expression was found in 38% of cases, CD3 in 33%, CD4 in 50%, CD5 in 94%, CD8 in 55%, and CD10 in 35%. Four presenting features were found to confer an increased risk of treatment failure: age greater than or equal to 15 years, L2 lymphoblast morphology, abnormal karyotype, and membrane CD3 expression. This study illustrates the heterogeneity of presentations of childhood T-cell ALL and suggests that the relative importance of risk factors in ALL differs according to immunophenotype and treatment strategy.     

Immunophenotype and clinical characteristics of CD45-negative and CD45-positive childhood acute lymphoblastic leukemia

 To evaluate the expression pattern of the leukocyte common antigen CD45 in acute leukemias and to investigate whether the lack of CD45 expression in childhood acute lymphoblastic leukemia (ALL) is associated with other immunophenotypic features and a distinct clinical behavior, we have carried out extensive immunophenotypic analyses of bone marrow and peripheral blood samples from 638 patients with childhood B-cell precursor (n=529) or T-lineage ALL (n=109). All 638 patients were enrolled in the German ALL-BFM 90 and ALL-BFM 95 trials. CD45 was detected on the surface of childhood ALL cells (cut-off ≥20% positive cells) in only 88.7% (n=566) of all cases. Among 529 patients with childhood B-cell precursor ALL, 12.9% (n=68) did not express CD45, compared with only 3.7% (n=4) of patients with childhood T-lineage ALL (p<0.001). In the B-cell precursor ALL subtypes, the highest frequency of CD45- cases (15.1%) was observed in common ALL (56/372) compared with only 7.2% in pro-B ALL (3/41) and 7.8% in pre-B ALL (9/116). Assessment of clinical parameters (age, organ enlargement, WBC, etc.) and event-free survival did not reveal significant differences between CD45- and CD45+ patients. Myeloid antigen coexpression was not correlated with CD45 expression. The mean percentage of antigen expression for CD34, CD10, TdT, CD22, and CD24 was significantly higher in children with CD45- B-cell precursor ALL than in those with CD45+ B-cell precursor ALL. In 28 patients with B-cell precursor ALL, cell cycle analyses of freshly isolated leukemic cells were performed with propidium iodide (PI) staining and flow-cytometric analysis. The percentage of cells in S-phase was inversely correlated to the percentage of CD45+ cells (r=-0.48, p<0.05). With two-parameter analysis of CD45-fluorescein isothiocyanate (FITC)- and PI-stained cells in nine patients with a percentage of CD45+ cells between 40 and 60%, two populations were distinguishable in a single patient. It was shown that the CD45- subpopulation had a higher percentage of cells in S-phase than the CD45+ subpopulation (10.7±4.0 vs. 2.7±1.8, p<0.007). We conclude that the lack of CD45 expression contributes to the identification of a distinct functional and immunological subgroup of B-cell precursor ALL, but that it has no significant impact on clinical behavior or on therapy outcome in childhood ALL. Received: February 18, 1998 / Accepted: May 26, 1998     

Impact of CD133 (AC133) and CD90 expression analysis for acute leukemia immunophenotyping

BACKGROUND AND OBJECTIVES: AC133 is a novel monoclonal antibody (Moab) reacting with a population of immature/primitive or granulo-monocytic committed CD34+ve cells. Up to now, only few studies with small numbers of cases have examined AC133 (recently designated CD133) expression in acute leukemia. To determine the value of this Moab for acute leukemia immunophenotyping, we investigated a large series of leukemic cell samples for their reactivity with Moab AC133. DESIGN AND METHODS. A total of 298 cell samples from patients with de novo acute myeloid leukemia (AML) (n=142), acute lymphoblastic leukemia (ALL) (n=119), CD34+ve biphenotypic acute leukemia (n=13), and CD34+ve CML blast crisis (=BC; 21 myeloid BC/3 lymphoid BC) were investigated by flow cytometry for Moab AC133 reactivity.CD133 expression was compared with CD90(Thy-1) expression, another CD34-associated antigen. RESULTS: Fifteen (5%) samples expressed CD90, whereas 114 (38%) samples were positive for Moab AC133 (20% cut-off level). No significant differences in CD133 and CD90 expression levels between AML and ALL were observed. In AML, but not ALL, CD133 was more often expressed in CD34+ve cases than in CD34-ve ones (p<0.00001). However, CD133 expression was not restricted to CD34+ve leukemic cells in individual cell samples. All 8 pro-B-ALL cell samples with 11q23-anomalies and MLL (mixed lineage leukemia) gene translocations were positive for CD133, whereas only 2 of 9 pro-B-ALL without MLL gene translocations expressed CD133 (p<0.002). In contrast, none of the 5 AML cell samples with a t(9;11) and MLL gene translocation reacted with Moab AC133. CD34+ve CML cells in myeloid BC were less often positive for CD133 than CD34+ve de novo AML cells (p<0.0001). INTERPRETATION AND CONCLUSIONS: CD133 and CD90 expression analysis is not helpful for lineage determination in acute leukemia immunophenotyping. However, MoabAC133 may be an informative marker for the detection and further characterization of immature AML cells, as well as pro-B-ALL cells with MLL gene translocations, by flow cytometry.     

Low expression of costimulatory molecules and mRNA for cytokines are important mechanisms of immunosuppression in acute lymphoblastic leukemia in children?

Mechanisms leading blasts of acute lymphoblastic leukemia to escape from immune surveillance are still unknown. Only few reports showed that ALL cells are inefficient antigen presenting cells. The aim of the study was to assess expression of critical costimulatory/adhesion molecules and mRNA for main pro- and anti-inflammatory cytokines in ALL cells. Children with B-cell precursor ALL (n=20) were prospectively enrolled into the study. Expression of costimulatory/adhesion molecules (CD1a, CD11c, CD40, CD54, CD80, CD83, CD86, CD123, HLA class I and II) was assessed by flow cytometry and mRNA for cytokines (IFN-gamma, IL-10, IL-4, TGF-beta) - with real-time PCR. Results: 1) high expression was observed for HLA I and II class, moderate for CD40, CD83, CD86 and low or no expression for CD80, CD54, CD1a, CD11c and CD123; 2) we found expression of mRNA for IFN-gamma, IL-10, IL-4 and TGF-beta in blasts cells (but not in all specimens). We noted relatively lower expression of all assessed cytokines comparing to T-cells obtained from healthy donors but interestingly expression for IL-10 was higher in normal B-cells than in blast cells, and IFN-gamma and IL-4 were not found in normal B-cells. In summary we suggest that ALL-blasts present low expression of costimulatory/adhesion molecules and mRNA for cytokines and this probably contribute to the absence of host T- cells stimulation to immune response.     

Improved outcome in adult B-cell acute lymphoblastic leukemia

A total of 68 adult patients with B-cell acute lymphoblastic leukemia (B-ALL) were treated in three consecutive adult multicenter ALL studies. The diagnosis of B-ALL was confirmed by L3 morphology and/or by surface immunoglobulin (Slg) expression with > 25% blast cell infiltration in the bone marrow (BM). They were characterized by male predominance (78%) and a median age of 34 years (15 to 65 y) with only 9% adolescents (15 to 20 y), but 28% elderly patients (50 to 65 y). The patients received either a conventional (N = 9) ALL treatment regimen (ALL study 01/81) or protocols adapted from childhood B-ALL with six short, intensive 5-day cycles, alternately A and B. In study B-NHL83 (N = 24) cycle A consisted of fractionated doses of cyclophosphamide 200 mg/m2 for 5 days, intermediate-dose methotrexate (IdM) 500 mg/m2 (24 hours), in addition to cytarabine (AraC), teniposide (VM26) and prednisone. Cycle B was similar except that AraC and VM26 were replaced by doxorubicin. Major changes in study B-NHL86 (N = 35) were replacement of cyclophosphamide by ifosphamide 800 mg/m2 for 5 days, an increase of IdM to high-dose, 1,500 mg/m2 (HdM) and the addition of vincristine. A cytoreductive pretreatment with cyclophosphamide 200 mg/m2, and prednisone 60 mg/m2, each for 5 days was recommended in study B-NHL83 for patients with high white blood cell (WBC) count (> 2,500/m2) or large tumor burden and was obligatory for all patients in study B-NHL86. Central nervous system (CNS) prophylaxis/treatment consisted of intrathecal methotrexate (MTX) therapy, later extended to the triple combination of MTX, AraC, and dexamethasone, and a CNS irradiation (24 Gy) after the second cycle. Compared with the ALL 01/81 study where all the patients died, results obtained with the pediatric protocols B-NHL83 and B-NHL86 were greatly improved. The complete remission (CR) rates increased from 44% to 63% and 74%, the probability of leukemia free survival (LFS) from 0% to 50% and 71% (P = .04), and the overall survival rates from 0% to 49% and 51% (P = .001). Toxicity, mostly hematotoxicity and mucositis, was severe but manageable. In both studies B-NHL83 and B-NHL86, almost all relapses occurred within 1 year. The time to relapse was different for BM, 92 days, and for isolated CNS and combined BM and CNS relapses, 190 days (P = .08). The overall CNS relapses changed from 50% to 57% and 17%, most probably attributable to the high-dose MTX and the triple intrathecal therapy. LFS in studies B-NHL83 and B-NHL86 was significantly influenced by the initial WBC count < or > 50,000/microL, LFS 71% versus 29% (P = .003) and hemoglobin value > or < 8 g/dL, LFS 67% versus 27% (P = .02). Initial CNS involvement had no adverse impact on the outcome. Elderly B- ALL patients (> 50 years) also benefited from this treatment with a CR rate of 56% and a LFS of 56%. It is concluded that this short intensive therapy with six cycles is effective in adult B-ALL. HdM and fractionated higher doses of cyclophosphamide or ifosphamide seem the two major components of treatment.     

RAC2, AEP, and ICAM1 expression are associated with CNS disease in a mouse model of pre-B childhood acute lymphoblastic leukemia

We developed a murine model of CNS disease to obtain a better understanding of the pathogenesis of CNS involvement in pre-B-cell acute lymphoblastic leukemia (ALL). Semiquantitative proteomic discovery-based approaches identified unique expression of asparaginyl endopeptidase (AEP), intercellular adhesion molecule 1 (ICAM1), and ras-related C3 botulinum toxin substrate 2 (RAC2), among others, in an invasive pre-B-cell line that produced CNS leukemia in NOD-SCID mice. Targeting RAC2 significantly inhibited in vitro invasion and delayed disease onset in mice. Induced expression of RAC2 in cell lines with low/absent expression of AEP and ICAM1 did not result in an invasive phenotype or murine CNS disease. Flow cytometric analysis identified an enriched population of blast cells expressing ICAM1/lymphocyte function associated antigen-1 (LFA-1)/CD70 in the CD10(+)/CD19(+) fraction of bone marrow aspirates obtained from relapsed compared with normal controls and those with primary disease. CD10(+)/CD19(+) fractions obtained from relapsed patients also express RAC2 and give rise to CNS disease in mice. Our data suggest that combinations of processes are involved in the pathogenesis of CNS disease in pre-B-cell ALL, support a model in which CNS disease occurs as a result of external invasion, and suggest that targeting the processes of adhesion and invasion unique to pre-B cells may prevent recurrences within the CNS.     

CD99 (MIC2) expression in paediatric B-lineage leukaemia/lymphoma reflects maturation-associated patterns of normal B-lymphopoiesis

We have recently shown that CD99 (MIC2) is differentially expressed during normal early B-cell development in the bone marrow (BM). Since immature B-cell precursors (BCP) are assumed to correspond to some extent to acute lymphoblastic leukaemia (ALL) and non-Hodgkin's lymphoma (NHL) cells with respect to patterns of phenotypic differentiation, we wondered whether the particular maturation-associated expression patterns of CD99 in the normal BCP stages were conserved also in malignant cells. Therefore we compared malignant and physiological B cells from paediatric ALL/NHL and normal BM samples with respect to CD99 expression using selective gating and semi-quantitative flow cytometry. Common-ALLs (n = 45) were similar to their corresponding, very immature BCPs (stage 1) in expressing very high levels of CD99. Most pre-B ALLs (n = 16) were also CD99hi and thus differed from the patterns found in normal cytoplasmic mu-chain+ (cmu+) pre-B cells (stage 2, CD99lo). In particular, we found that those pre-B-ALL cases which were CD34+ also showed higher CD99 expression than the CD34- cases. This prompted us to investigate the levels of CD99 in those rare normal BCPs which also coexpress CD34 and cmu; these cells, which are transitory from stage 1 to stage 2, were found also CD99hi, thus precisely reflecting the patterns of CD34+ pre-B ALLs. The blasts of Burkitt-type B-cell ALL/NHL samples (n = 13) expressed considerably less CD99, similarly to the more differentiated BCP stages 2 (cmu+) and 3 (surface mu-chain+). In summary, we found that paediatric B-lineage malignancies display remarkable synchrony regarding the levels of CD99 expression compared to their putative normal counterparts.     

The generation of immunocompetent dendritic cells from CD34+ acute myeloid or lymphoid leukemia cells

The ability of CD34+ leukemic cells to differentiate to dendritic cells (DCs) was investigated in 18 acute myeloid leukemia (AML) and 4 lymphoid leukemia (ALL) patients. The generation of DCs was determined by the expression of DC-associated CD1a or CD83 (more than 30%) with costimulatory molecules, by CD80 antigens (>20%), and by the exhibition of allostimulatory activity. In the AML patients, allostimulatory mature DCs were generated from 3 of 9 M0 or M1, 2 of 5 M2,2 of 4 M4 or M5, and 3 of 4 ALL (L2) cases. In total, DCs were more efficiently induced from cases expressing over 75% of CD34+ among whole bone marrow mononuclear cells (8 of 12), compared with those under 75% (2 of 10; P < .05). B-cell (CD19), natural killer (NK)-cell (CD56), or T-cell (CD7) lineage markers, which were aberrantly expressed on the blasts, were rarely found on leukemic DCs at the end of the culture period, and myeloid (CD13, CD33), not lymphoid (CD10), markers were shown on ALL-derived DCs. In Philadelphia chromosome-positive ALL or AML patients with t (8;21), DCs were confirmed to be of leukemic origin by fluorescence in situ hybridization analysis.     

Transplantation of highly purified CD34+ progenitor cells from unrelated donors in pediatric leukemia

Unrelated donors are commonly used for hematopoietic stem cell transplants, but graft-versus-host disease (GVHD) is a major problem. We investigated whether transplantation of purified mobilized peripheral-blood CD34(+) stem cells from unrelated donors would prevent acute and chronic GVHD in pediatric patients with leukemia and avert the need for pharmacologic immunosuppression. Thirty-one pediatric patients with acute lymphoblastic leukemia (ALL, n = 16), acute myeloid (n = 7), chronic myeloid (n = 6), or juvenile myelomonocytic leukemia (n = 2) underwent transplantation. The median purity of CD34(+) cells after positive magnet-activated cell sorting was 98.5%. Patients received a median of 8.0 x 10(6) CD34(+) cells and 6 x 10(3) CD3(+) T lymphocytes per kilogram, with no posttransplantation pharmacologic immunosuppression. Primary acute GVHD > or = grade II was seen in only 10% of patients (n = 3) and occurred only after human herpesvirus 6 (HHV 6) infection. Two patients had limited chronic GVHD. Engraftment occurred in all patients (primary engraftment, n = 26; engraftment after reconditioning, n = 5). The 2-year survival estimate was 38% for all patients and 63% for patients with ALL in complete remission. Patients with myeloid malignancies had a poor outcome. In comparison to a historical control group who received unmanipulated bone marrow, our patients had a lower incidence of GVHD (P <.001). No difference was observed in the probability of relapse or survival. Study patients with ALL in remission showed a trend toward better survival (P =.07). Transplantation of purified peripheral-blood CD34(+) cells from unrelated donors effectively minimizes GVHD and may be a good therapeutic option for patients with relapsed ALL.     

CD2 antigen expression on leukemic cells as a predictor of event-free survival after chemotherapy for T-lineage acute lymphoblastic leukemia: a Children's Cancer Group study

We examined the prognostic impact of CD2 antigen expression for 651 patients with T-lineage acute lymphoblastic leukemia (ALL), who were enrolled in front-line Childrens Cancer Group treatment studies between 1983 and 1994. There was a statistically significant correlation between the CD2 antigen positive leukemic cell content of bone marrow and probability of remaining in bone marrow remission, as well as overall event-free survival (EFS) (P = .0003 and P = .002, log-rank tests for linear trend). When compared with patients with the highest CD2 expression level (> 75% positivity), the life table relative event rate (RER) was 1.22 for patients with intermediate range CD2 expression level (30% to 75% positivity) and 1.81 for "CD2-negative" patients (< 30% positivity). At 6 years postdiagnosis, the EFS estimates for the three CD2 expression groups (low positivity to high positivity) were 52.8%, 65.5%, and 71.9%, respectively. CD2 expression remained a significant predictor of EFS after adjustment for the effects of other covariates by multivariate regression, with a RER of 1.47 for CD2- negative patients (P = .04). Analysis of T-lineage ALL patients shows a significant separation in EFS after adjustment for the National Cancer Institute (NCI) age and white blood cell (WBC) criteria for standard and high-risk ALL (P = .002, RER = 1.67). The determination of CD2 expression on leukemic cells helped identify patients with the better and poorer prognoses in both of these risk group subsets. For standard risk T-lineage ALL, CD2-negative patients had a worse outcome (P = .0007, RER = 2.92) with an estimated 5-year EFS of 55.9% as compared with 78.3% for the CD2-positive patients. Thus, CD2 negativity in standard risk T-lineage ALL identified a group of patients who had a worse outcome than high-risk T-lineage ALL patients who were CD2 positive. The percentage of CD2 antigen positive leukemic cells from T- lineage ALL patients is a powerful predictor of EFS after chemotherapy. This prognostic relationship is the first instance in which a biological marker in T-lineage ALL has been unequivocally linked to treatment outcome.     

Rare adult acute lymphocytic leukemia with CD56 expression in the ECOG experience shows unexpected phenotypic and genotypic heterogeneity

Expression of CD56, a marker of natural killer (NK) cells, in acute lymphocytic leukemia (ALL) is rare and, to date, has been described only in non-B lineage ALL. Among 194 patients with CD56 analysis on the ongoing Eastern Cooperative Oncology Group (ECOG) ALL trial, E2993, 6 cases of CD56+ ALL were found (3.1%) with a median of 95% of blast cells expressing CD56, compared with a median of 1% of blast cells in CD56- ALL (P = 0.0001). FAB-L2 characteristics dominated, without granulation. Blast cells from four CD56+ patients expressed T-cell antigens at variable levels of maturation. A clonal rearrangement of the T-cell receptor beta (TCRbeta) gene was detected only in one patient. TCRbeta variable gene usage studies in this and one other CD56+ ALL patient demonstrated a significantly perturbed usage pattern in both patients when compared with control lymphocytes. The two remaining cases typed as early pre-B ALL (CD19+, CD10+), with one case co-expressing CD7. Cytogenetically, 4 patients were normal, 1 complex abnormal, and 1 Philadelphia chromosome positive. Epstein-Barr virus (EBV) sequences were detected in one T- and both B-lymphoid cases. Our data suggest that CD56 is expressed at a precursor stage common to the T- and the B-cell lineage.     

Philadelphia chromosome-positive acute lymphoblastic leukemia in Taiwan

From January 1986 to December 1998, 26 (23%) of 114 acute lymphoblastic leukemia (ALL) patients older than 15 years were found to have Philadelphia (Ph) chromosome. They accounted for 28% (26 of 94) of the patients with B-lineage ALL. Compared with the other 88 ALL patients, the leukemic cells from all but one Ph-positive ALL patients were early pre-B cells with a higher rate of CD34 expression (92% vs 50%, P<0.05). At presentation, the Ph-positive adult ALL patients had higher circulating blasts (mean 21.4 vs 3.66x10(4)/microl, P<0.05) and lower initial platelet counts (mean 4.47 vs 7.34x10(4)/microl, P<0.01) and showed a trend toward higher frequency of initial central nerve system (CNS) involvement (25% vs 11%, P=0.098) than the others. Among patients with adequate chemotherapy, 16 (64%) of the 25 Ph-positive ALL patients achieved complete remission (CR), an incidence marginally lower than that of Ph-negative ALL (62 of 76, 82%, P=0.06) and significantly lower than that of Ph-negative B-lineage ALL (50 of 58, 86%, P=0.0037). However, all patients relapsed except for those who received allogeneic hemopoietic stem cell transplantation (allo-HSCT). The probabilities of 5-year continuous CR and 5-year survival for Ph-positive adult ALL patients were significantly inferior to those for Ph-negative adult ALL patients (0% vs 12%, P=0.0001, and 7% vs 19%, P=0.047, respectively), and those for Ph-negative B-lineage ALL (0% vs 14%, P<0.0001, and 7% vs 23%, P=0.002, respectively). Prognostic factors were analyzed among the Ph-positive ALL patients including the 26 adults mentioned above and an additional 11 children. No clinical or biological characteristics such as age, sex, initial circulating blast count, extramedullary involvement, or CD34 expression had an impact on the disease outcome. Allo-HSCT in first CR may improve the probability of 5-year continuous CR (100% vs. 0% for those without allo-HSCT, P=0.0091) although only three patients received it in this study. In conclusion, Ph-positive ALL patients tended to have a poor prognosis, regardless of whether other possible risk factors were present or not. Aggressive treatment, such as high-dose chemotherapy with allo-HSCT, should be considered for these patients to improve survival.