Modulation of nitrosourea resistance in myeloid leukemias

Drug resistance in myeloid leukemias may be mediated by an increased capacity to repair chemotherapy-induced DNA damage. Some tumor cell lines that are resistant to nitrosoureas contain the DNA repair protein O6-alkylguanine-DNA alkyltransferase (alkyltransferase). This protects cells by removing cytotoxic, nitrosourea-induced O6-alkylguanine adducts. We measured the level of alkyltransferase activity in myeloid leukemic cells freshly obtained from patients to determine whether the alkyltransferase was an important factor in nitrosourea resistance in these cells and whether inactivation of this protein could sensitize leukemic cells to nitrosoureas. Myeloid leukemic cells from patients with acute nonlymphocytic leukemia and chronic myelogenous leukemia had higher levels of alkyltransferase than did myeloid precursors from normal donors (P less than .01). This difference did not appear to be due to the state of differentiation of the leukemic or normal cells. To show that this repair protein mediated nitrosourea resistance in leukemic cells, cells were treated with the modified base O6- methylguanine to selectively and irreversibly inactivate the alkyltransferase and then exposed to 1,3-bis (2-chloroethyl)-1- nitrosourea (BCNU). An 18-hour incubation in 0.5 mmol/L O6- methylguanine caused an 87% +/- 3.6% decrease in alkyltransferase activity in leukemic cells and a 73% +/- 8.6% decrease in normal myeloid precursors. After treatment with O6-methylguanine, clonogenic leukemic cells from ten different donors became much more sensitive to BCNU, with a decrease in the dose needed to reduce colony survival by 50% (LD50) of 6.3 +/- 1.4-fold. A lesser effect was seen on CFU-GM, BFU- E, and CFU-GEM where the LD50 decreased two- to threefold. These studies show that nitrosourea resistance in myeloid leukemic cells can be abrogated by inactivation of the DNA repair protein O6-alkylguanine- DNA alkyltransferase. This method of biochemical modulation of DNA repair will sensitize leukemic cells to nitrosoureas in vitro and has the potential of increasing the therapeutic index of nitrosoureas in this disease.     

Endothelial prostacyclin production is a late event in granulocyte migration into bovine pulmonary artery intimal explants

Whether migration of granulocytes across pulmonary vascular endothelium in the absence of structural evidence of endothelial injury causes increased production of thromboxane or prostacyclin is not known. Using bovine pulmonary artery intimal explants mounted in Boyden chambers and homologous separated granulocytes, concentrations of thromboxane B2 and 6-keto-PGF1 alpha in the upper-well fluid were measured by radioimmunoassay over a three-hour period under the following conditions: (1) granulocyte chemotaxis (zymosan-activated plasma in the lower well, granulocytes in the upper well); (2) unstimulated granulocyte migration (serum or plasma in the lower well, granulocytes in the upper well); (3) granulocyte activation without migration (zymosan-activated plasma and granulocytes in the upper well); (4) granulocyte chemotaxis in the absence of endothelium (identical to condition 1 above except that endothelium was scraped from the explant surface); and (5) explants incubated in the absence of granulocytes. Minimal increases in thromboxane B2 concentrations in upper-well fluid occurred under all conditions. In contrast, granulocyte chemotaxis was accompanied by large increases in concentrations of 6-keto-PGF1 alpha evident by two hours of incubation and increasing markedly by three hours, to 524.3 +/- 69.0 ng/mL (m +/- SEM). Unstimulated migration of granulocytes toward serum or plasma and granulocyte activation without migration were accompanied, at three hours, by more modest increases in 6-keto-PGF1 alpha (296.5 +/- 46.4; 128.0 +/- 38.6, and 236.7 +/- 47.0 ng/mL, respectively) and, in the absence of granulocytes or in the absence of endothelium, only minimal increases in this prostacyclin metabolite occurred (137.2 +/- 16.9 and 53.9 +/- 12.6 ng/mL, respectively). The large rises in prostacyclin metabolite occurred at a time when the majority of granulocytes had migrated through the endothelial layer rather than during their adherence or transendothelial passage. We conclude that chemotaxis of granulocytes through pulmonary vascular endothelium causes endothelial production of large amounts of prostacyclin, but this occurs late in the chemotactic process, after granulocytes have transversed the endothelium.     

Immature human cord blood progenitors engraft and proliferate to high levels in severe combined immunodeficient mice

Unseparated or Ficoll-Hypaque (Pharmacia, Piscataway, NJ)--fractionated human cord blood cells were transplanted into sublethally irradiated severe combined immunodeficient (SCID) mice. High levels of multilineage engraftment, including myeloid and lymphoid lineages, were obtained with 80% of the donor samples as assessed by DNA analysis, fluorescence-activated cell sorting (FACS), and morphology. In contrast to previous and concurrent studies with adult human bone marrow (BM), treatment with human cytokines was not required to establish high-level human cell engraftment, suggesting that neonatal cells either respond differently to the murine microenvironment or they provide their own cytokines in a paracrine fashion. Committed and multipotential myelo- erythroid progenitors were detected using in vitro colony assays and FACS analysis of the murine BM showed the presence of immature CD34+ cells. In addition, human hematopoiesis was maintained for at least 14 weeks providing further evidence that immature hematopoietic precursors had engrafted the murine BM. This in vivo model for human cord blood- derived hematopoiesis will be useful to gain new insights into the biology of neonatal hematopoietic cells and to evaluate their role in gene therapy. There is growing evidence that there are ontogeny-related changes in immature human hematopoietic cells, and therefore, the animal models we have developed for adult and neonatal human hematopoiesis provide useful tools to evaluate these changes in vivo.     

Relationship of the hematopoietic cycle to the combined activities of a stimulator and a short-lived inhibitor

Substances that circulate in the blood following drug-induced bone marrow aplasia produce a biphasic curve of DNA synthesis in cells in liquid and semisolid cultures which reflect relative concentrations of these growth regulators of hematopoiesis. This net effect magnified by induced marrow failure in human, rat, and mouse is a sinusoidal curve that is the reciprocal of the WBC. Generated in the bone marrow, humoral stimulating activity (HSA) produces peak growth of colonies in agar (CFU-GM) during the proliferative phase of bone marrow recovery, whereas humoral inhibitory activity (HIA), present at the time of marrow maturation, suppresses colony growth. Split femurs from rats given cyclophosphamide (CY) and killed at regular intervals condition media that affect DNA synthesis in a fashion similar to that of HSA-HIA in the rats' sera. In Dexter cultures, HSA is derived from the adherent rather than the hematopoietic cell, whereas HIA is produced by that nonadherent cell. Animals treated with a lethal dose of busulfan (BU) produce large amounts of HSA in vivo until death. Those transplanted with adherent bone marrow cells depleted of hematopoietic cells on day 1 after BU also die, whereas those given nonadherent marrow cells survive and generate HIA at the time of bone marrow recovery. HSA and media conditioned by bone marrow stromal cells causes an increase in spleen colonies (CFU-S), as does HIA. These studies support the contention that the net effect of putative regulators of hematopoiesis, measured in the drug-perturbed state, consist of a constantly present stimulator emanating from bone marrow stroma, and a variable inhibitor produced by maturing hematopoietic bone marrow cells.     

Significance of Ph1-negative marrow cells in Ph1-positive chronic granulocytic leukemia

Fifty-six of 195 Ph1-positive patients with chronic granulocytic leukemia were found to have Ph1-negative metaphases in marrow aspirates on one or more occasions. In 22 cases, Ph1-negative cells were found prior to initiation of antileukemic therapy. Five patients were in the blastic stage of the disease when Ph1-negative mitoses were seen. The finding of Ph1-negative cells appeared to be related principally to short duration of CGL and to administration of antileukemic therapy (conventional agents and doses, in most cases). Ph1-negative cells were usually not found more than 2 yr after the diagnosis of leukemia, but in a few cases, they were seen as long as 5-10 yr after diagnosis. Only a minority of metaphases analyzed were Ph1-negative, except in the case of 6 patients who transiently had 50% or more Ph1-negative cells after antileukemic therapy. The presence of Ph1-negative cells in marrow was not associated with any survival advantage in this series.     

Prognostic significance of additional cytogenetic abnormalities at diagnosis of Philadelphia chromosome-positive chronic granulocytic leukemia

Of 661 patients with Philadelphia chromosome (Ph)-positive, nonblastic chronic granulocytic leukemia, 58 had cytogenetic abnormalities in addition to the Ph at the time of diagnosis. Twenty patients had reduplication of the Ph in one or more metaphases. Twenty-one patients with a single Ph exhibited hyperdiploidy in one or more metaphases. Eleven patients had two or more hypodiploid metaphases as their only numerical abnormality. The remaining six patients had a variety of abnormalities. Many patients had more than one type of abnormality. Survival of patients in the different subgroups was similar, but these 58 patients had a shorter course than the 603 patients without additional cytogenetic abnormalities (P less than .02). Survival curves for the two populations did not diverge until the 2-year point, after which the annual death rate among patients with additional cytogenetic abnormalities was approximately 40% higher than that of patients without such abnormalities. The two populations had similar relative risk values according to a hazard ratio formula previously described by the International CGL Prognosis Study Group. Thus, they would have been expected to have essentially identical survival curves. We conclude that the presence of additional cytogenetic abnormalities at the time of diagnosis constitutes an independently significant prognostic feature with an unusually delayed influence on survival.     

Focus-forming ability and surface markers of hamster-human malignant lymphoma hybrids

We have examined the properties of hybrid cells formed by polyethylene glycol-mediated fusion of the GRC+L-73 line of Chinese hamster ovary (CHO) cells with peripheral blood cells from patients with chronic lymphocytic leukemia (CLL) or with bone marrow cells from patients with malignant lymphoma. The results indicate that hybrid cells can be detected by their ability to form "foci" of characteristic morphology in the presence of a monolayer of parental CHO cells and that clones isolated from such foci express aspects of the differentiation status, as detected by immunologic markers, of the human parental cells.     

The sensitivity to cytosine arabinoside of the blast progenitors of acute myeloblastic leukemia

Two culture methods are available for the study of the blast cells of acute myeloblastic leukemia (AML). One is an assay for clonogenic precursors; it depends on their ability to form blast colonies in culture in the presence of methylcellulose and suitable growth factors. The other assesses the growth of blast cells in suspension culture, where growth is measured by increasing numbers of clonogenic cells. We have compared the two methods as assays for the cytotoxic effects of the chemotherapeutic drug cytosine arabinoside (Ara-C). Marked patient- to-patient variation was found using either method; however, the slopes of the dose-response curves were usually greater when cells were exposed to drug in suspension rather than in methylcellulose. Control experiments showed that the difference could not be explained by drug carry-over from the suspension cultures to the methylcellulose plates when clonogenic cells in the suspensions were assessed. Further, the survival curves for Adriamycin were very similar, regardless of which assay was used. No correlation was found between D10 Ara-C values measured in suspension or in methylcellulose. However, a significant association with outcome was found between D10 Ara-C in suspension and response to treatment with a regimen in which Ara-C was the only chemotherapeutic agent used. No such association was detected when the D10 values obtained with the clonogenic assay were compared with outcome for the same group of 15 patients. Finally, a feasibility experiment was performed in which blast cells were exposed to Ara-C repeatedly during exponential growth over 238 days. A dose-related inhibition of growth was observed; no evidence was seen of emerging drug-resistant cells. Nor did the morphology of the cells change as a result of drug exposure. We conclude that drug sensitivities of AML blast cells in culture are dependent on measurement methods, even when techniques affecting cell proliferation are compared. Measurements of drug sensitivity in culture may best be interpreted when the bases of the assay systems are understood.     

Frequency and prognostic significance of HRX rearrangements in infant acute lymphoblastic leukemia: a Pediatric Oncology Group study

Chromosome band 11q23, the location of the HRX gene, is a site of recurrent translocations in human malignancies. Infants with acute lymphoblastic leukemia (ALL) commonly have 11q23 translocations and have an especially poor prognosis despite intensive chemotherapy. We analyzed 96 cases of infant ALL treated on three consecutive Pediatric Oncology Group protocols to determine the frequency and prognostic significance of molecular rearrangements of HRX. Overall, 78 cases (81%) had HRX rearrangements detected by Southern blot analysis performed with a single HRX cDNA probe, whereas 18 cases (19%) had germline HRX. Of the 78 cases with HRX rearrangements, only 50 had abnormalities of 11q23 detected cytogenetically. Molecular abnormalities of HRX were associated with early treatment failure and a very poor outcome. Estimated event-free survival for patients with HRX rearrangements was 19% (SE, 7%) at 3 years, compared with 46% (SE, 17%) for patients with germline HRX (P = .033 by the two-sided logrank test). Therefore, infants with ALL and molecular abnormalities of HRX represent a group with an extremely high rate of failure who clearly need innovative or experimental treatment. Furthermore, cytogenetic analysis alone failed to detected 36% of HRX rearrangements, suggesting that molecular analysis be performed on all infants with ALL to identify this group of high-risk patients.     

Asynchronous antigen expression in B lineage acute lymphoblastic leukemia

Cell surface phenotypes of 113 B lineage acute lymphocytic leukemia (ALL) cases, defined by the presence of HLA-DR and at least one B-cell- specific antigen (either CD19, CD20, or CD22), were compared with antigen-defined stages of normal B lymphocyte development. The cases were first evaluated for expression of HLA-DR, CD19, CD34, CD10, CD20, and CD22 by indirect one-color immunofluorescence. Pairwise comparisons of cell surface marker expression were performed for each leukemic sample: no correlations were observed for paired antigen expression on the leukemic samples using antigens expressed either early or late during normal B lymphoid development. Complete immunophenotypes of the cases were then compared with normal B-cell developmental stages. Sixteen different complete immunophenotypes were observed on the leukemias that were not found in normal marrow; at least 78% of the cases demonstrated such "asynchronous" combinations of B lymphoid- associated differentiation antigens. Several samples were subsequently studied by two-color immunofluorescence, and the presence of doubly labeled cells with "asynchronous" antigen combinations was confirmed. These results indicate that the majority of B lineage leukemias exhibit "developmental asynchrony," as compared with normal marrow B cells. The data further suggest that ALL cases do not accurately represent cells arrested at the stage where the leukemogenic event occurred. Rather, ALL appears to be a disease in which there may be maturation of leukemic blasts; but this maturation is "asynchronous" when compared with the normal developmental process.