Risk of relapse after anti-PD1 discontinuation in patients with Hodgkin lymphoma.

INTRODUCTION: Patients with relapsed/refractory Hodgkin lymphoma (R/R HL) experience high response rates upon anti-PD1 therapy. In these patients, the optimal duration of treatment and the risk of relapse after anti-PD1 discontinuation are unknown. METHODS: We retrospectively analyzed patients with R/R HL who responded to anti-PD1 monotherapy and discontinued the treatment either because of unacceptable toxicity or prolonged remission. A machine learning algorithm based on 17 candidate variables was trained and validated to predict progression-free survival (PFS) landmarked at the time of discontinuation of anti-PD1 therapy. RESULTS: Forty patients from 14 centers were randomly assigned to training (n = 25) and validation (n = 15) sets. At the time of anti-PD1 discontinuation, patients had received treatment for a median duration of 11.2 (range, 0-time to best response was not statistically significant in discriminating patients with PFS lesser or greater than 12 months). Considering PFS status as a binary variable (alive or dead) at a specific time point (12 months) is convenient, intuitive and allows for comparing the value of potential predicting variables in these two groups of patients. Nonetheless, this approach has two drawbacks: first, it binarizes outcome; second, it excludes patients alive with a time to last follow up lesser 12 months. Therefore, it is less powerful to demonstrate statistically significant association with PFS even if it exists 5 months. Patients discontinued anti-PD1 treatment either because of prolonged remission (N = 27, 67.5%) or unacceptable toxicity (N = 13, 32.5%). Most patients were in CR (N = 35, 87.5%) at the time of anti-PD1 discontinuation. In the training set, the machine learning algorithm identified that the most important variables to predict PFS were patients' age, time to best response, and presence or absence of CR. The performance observed in the training set was validated in the validation set. CONCLUSION: In this pilot, proof of concept study using a machine learning algorithm, we identified biomarkers capable of predicting the risk of relapse after anti-PD1 discontinuation (age, time to best response, quality of response). Once confirmed, these simple biomarkers will represent useful tools to guide the management of these patients.

High-dose sequential chemotherapy (HDS) versus PEB chemotherapy as first-line treatment of patients with poor prognosis germ-cell tumors: mature results of an Italian randomized phase II study.

BACKGROUND: In the late 1990s, the use of high-dose chemotherapy (HDCT) and stem-cell rescue held promise for patients with advanced and poor prognosis germ-cell tumors (GCT). We started a randomized phase II trial to assess the efficacy of sequential HDCT compared with cisplatin, etoposide, and bleomycin (PEB). PATIENTS AND METHODS: Patients were randomly assigned to receive four cycles of PEB every 3 weeks or two cycles of PEB followed by a high-dose sequence (HDS) comprising HD-cyclophosphamide (7.0 g/m(2)), 2 courses of cisplatin and HD-etoposide (2.4 g/m(2)) with stem-cell support, and a single course of HD-carboplatin [area under the curve (AUC) 27 mg/ml × min] with autologous stem-cell transplant. Postchemotherapy surgery was planned on responding residual disease in both arms. The primary end point was progression-free survival (PFS). The study was designed to detect a 30% improvement of 5-year PFS (from 40% to 70%), with 80% power and two-sided α at 5%. RESULTS: From December 1996 to March 2007, 85 patients were randomized: 43 in PEB and 42 in HDS arm. Median follow-up was 114.2 months [interquartile range (IQR): 87.7-165.8]. Complete or partial response with normal markers (PRm-) were obtained in 28 (65.1%) and 29 (69.1%) patients, respectively. Five-year PFS was 55.8% [95% confidence interval (CI) 42.8-72.8] and 54.8% (95% CI 41.6%-72.1%) in PEB and HDS arm, respectively (log-rank test P = 0.726). Five-year overall survival was 62.8% (95% CI 49.9-79.0) and 59.3% (95% CI 46.1-76.3). One toxic death (PEB arm) was recorded. CONCLUSIONS: The study failed to meet the primary end point. Furthermore, survival estimates of conventional-dose chemotherapy higher than expected should be accounted for and will likely limit further improvements in the first-line setting. CLINICALTRIALS.GOV: NCT02161692.

Infections after T-replete haploidentical transplantation and high-dose cyclophosphamide as graft-versus-host disease prophylaxis.

BACKGROUND: Recently, a platform of T-cell replete haploidentical hematopoietic stem cell transplantation (haplo-HSCT) using post-transplant cyclophosphamide (Cy) has shown high reproducibility and acceptable safety profile. METHOD: This prospective cohort analysis allowed us to collect data on infections among 70 consecutive recipients of haplo-HSCT affected by various hematologic malignancies. RESULTS: After a median follow-up of 23 months, cumulative incidence of viral infections was 70% (95% confidence interval [CI] 59-81) at 100 days and 77% (95% CI 67-87) at 1 year; 35 of 65 patients at risk had CMV reactivation (54%) and the rate of polyomavirus-virus-associated cystitis was 19% (13/70). Cumulative incidence of bacterial and fungal infections at 1 year were 63% (95% CI 51-75) and 12% (95% CI 4-19), respectively. Of note, only 1 invasive fungal infection occurred beyond 1 year after transplant (day +739). CONCLUSION: In conclusion, despite a high rate of viral infections in the early period, present data suggest a satisfactory infectious profile after T-cell replete haplo-HSCT using post-transplant Cy. These results may help clinicians to improve both prophylactic and therapeutic antimicrobial strategies in this emerging haploidentical setting.

Recurrence-specific supervised graph clustering for subtyping Hodgkin Lymphoma radiomic phenotypes.

The prediction at baseline of patients at high risk for therapy failure or recurrence would significantly impact on Hodgkin Lymphoma patients treatment, informing clinical practice. Current literature is extensively searching insights in radiomics, a promising framework for high-throughput imaging feature extraction, to derive biomarkers and quantitative prognostic factors from images. However, existing studies are limited by intrinsic radiomic limitations, high dimensionality among others. We propose an exhaustive patient representation and a recurrence-specific multi-view supervised clustering algorithm for estimating patient-to-patient similarity graph and learning recurrence probability. We stratified patients in two risk classes and characterize each group in terms of clinical variables.

Absorbed dose and biologically effective dose in patients with high-risk non-Hodgkin's lymphoma treated with high-activity myeloablative 90Y-ibritumomab tiuxetan (Zevalin).

PURPOSE: The aim of this study was to carry out two different dose estimation approaches in patients with non-Hodgkin's lymphoma (NHL) treated with a myeloablative amount of (90)Y-labelled ibritumomab tiuxetan (Zevalin(R)) in an open-label dose escalation study. METHODS: Twenty-seven patients with relapsed/refractory or de novo high-risk NHL receiving one myeloablative dose of (90)Y-ibritumomab tiuxetan followed by tandem stem cell reinfusion were evaluated for dose estimate. The injected activity was 30 MBq/kg in 12 patients and 45 MBq/kg in 15 patients. Dose estimation was performed 1 week prior to (90)Y-ibritumomab tiuxetan by injection of (111)In-ibritumomab tiuxetan (median activity: 200 MBq). The absorbed dose (D) and the biologically effective dose (BED) were calculated. RESULTS: The absorbed doses per unit activity (Gy/GBq) were [median (range)]: heart wall 4.6 (2.5-9.7), kidneys 5.1 (2.8-10.5), liver 6.1 (3.9-10.4), lungs 2.9 (1.5-6.8), red marrow 1.0 (0.5-1.7), spleen 7.0 (1.5-14.4) and testes 4.9 (2.9-16.7). The absorbed dose (Gy) for the 15 patients treated with 45 MBq/kg were: heart wall 17.0 (8.7-25.4), kidneys 17.1 (7.9-22.4), liver 20.8 (15.4-28.3), lungs 8.1 (5.4-11.4), red marrow 3.1 (2.0-4.0), spleen 26.2 (17.0-35.6) and testes 17.3 (9.0-28.4). At the highest activities the acute haematological toxicity was mild or moderate and of very short duration, and it was independent of the red marrow absorbed dose. No secondary malignancy or treatment-related myelodysplastic syndrome was observed. No non-haematological toxicity (liver, kidney, lung) was observed during a follow-up period of 24-48 months. CONCLUSION: The use of 45 MBq/kg of (90)Y-ibritumomab tiuxetan in association with stem cell autografting resulted in patients being free of toxicity in non-haematological organs. These clinical findings were in complete agreement with our dose estimations, considering both organ doses and BED values.

Haploidentical transplantation with post-infusion cyclophosphamide in advanced Hodgkin lymphoma.

We investigated the use of haploidentical hematopoietic stem cell transplantation (haplo-HSCT) in the treatment of advanced Hodgkin lymphoma (HL). Sixty-two consecutive HL patients underwent haplo-HSCT. Unmanipulated stem cells and post-transplant cyclophosphamide were given to all patients as GVHD prophylaxis. At 100 days, the cumulative incidence of grades 2-3 and grades 3-4 acute GVHD was 23% and 4%, respectively. The chronic GVHD (cGVHD) cumulative incidence was 16%, with one patient experiencing severe cGVHD. The 3-year OS, PFS, relapse rates and 1-year non-relapse mortality (NRM) were 63%, 59%, 21% and 20%, respectively. Uncontrolled disease status and high hematopoietic cell transplantation comorbidity index (HCT-CI) were associated with lower OS, whereas PBSC was an independent protective factor. Uncontrolled disease and HCT-CI >2 was predictive for NRM. Finally, disease status other than CR was predictive of relapse. In conclusion, haplo-HSCT is a valid treatment in advanced HL, offering excellent rates of survival and acceptable toxicities.

Marked telomere shortening in mobilized peripheral blood progenitor cells (PBPC) following two tightly spaced high-dose chemotherapy courses with G-CSF.

The purpose of the study was to compare telomere length (TL) in peripheral blood progenitor cells (PBPC) collected after two tightly spaced high-dose (hd) chemotherapy courses. We assessed 37 previously untreated lymphoma patients undergoing a hd-chemotherapy program with autografting. They sequentially received hd-cyclophosphamide (CY) and hd-Ara-C, both followed by PBPC harvesting. Both post-CY and post-Ara-C harvests were assessed for TL by Southern blot analysis. In 12 patients, the assay was also performed on purified CD34+ cells. All patients displayed high PBPC mobilization following both hd-CY and hd-Ara-C. In all but one patient, TL was shorter in PBPC collected after Ara-C compared to CY: 7226bp (range: 4135-9852) vs 8282 bp (range 4895-14860) (P < 0.0001). This result was confirmed on CD34+ cells. Platelet recovery in patients receiving post-Ara-C PBPC was significantly slower compared to those receiving post-CY PBPC. In conclusion, (i) administration of tightly spaced hd-chemotherapy courses induces marked telomere shortening on harvested PBPC; (ii) engraftment kinetics seem slower, with delayed platelet recovery, in patients autografted with PBPC suffering marked TL erosion; (iii) long-term follow-up is required to verify whether PBPC with shortened telomeres display defective engraftment stability and/or risk of secondary leukemia; (iv) TL evaluation is advisable whenever new mobilization procedures are developed.

Selective expression and constitutive phosphorylation of SHC proteins [corrected] in the CD34+ fraction of chronic myelogenous leukemias.

The BCR/ABL fusion protein is a constitutively active tyrosine kinase that is responsible for the pathogenesis of chronic myelogenous leukemia (CML). Clinically, CML is characterized by a chronic phase (CP) that eventually terminates into a blast crisis (BC). BC transformation is associated with accumulation of CD34+ blasts. We investigated the expression and phosphorylation of Src-homology-2 and collagen-homology domains (SHC) [corrected] proteins in subpopulations of CML primary cells. Shc polypeptides are tyrosine kinase substrates that are constitutively tyrosine-phosphorylated in continuous cell lines of CML origin. High levels of Shc expression were found in the CD34+ cells from CML-BC, CML-CP and normal bone marrow. In contrast, CD34- fractions from CML-CP and normal bone marrow expressed low levels of p46Shc. Shc proteins were constitutively phosphorylated in the CD34+ fractions from CML cells (both CP and BC), but not in normal CD34+ cells. These data bear implications for the role of Shc in normal hemopoiesis and CML leukemogenesis: (a) dramatic changes of Shc expression during terminal differentiation of hemopoietic cells adds a further level of regulation to the signal transduction function of Shc; and (b) constitutive Shc tyrosine-phosphorylation in the rare CD34+ cells of CML-CP might contribute to the selection of this subpopulation during the blast crisis transformation of CMLs.

Induction of apoptosis by arachidonic acid in chronic myeloid leukemia cells.

The hallmark of chronic myeloid leukemia (CML) is the presence of the bcr-abl oncogene, which is associated with transforming ability and an intrinsic resistance to induction of apoptosis by genotoxic agents. Arachidonic acid (AA), a biologically active fatty acid, plays a crucial role as a mediator of signaling pathways involved in cell proliferation and survival. In this study, we investigated the potential role of AA as a proapoptotic agent in CML. Pretreatment of human CML isolated progenitor cells with AA (100 microM for 18 h) induced 71-75% inhibition of in vitro colony formation of granulocyte-macrophage colony-forming units, multilineage colony-forming units, and erythroid burst-forming units. This inhibition was significantly greater than the effect on normal progenitor cells (19-39% growth inhibition of erythroid burst-forming units, multilineage colony-forming units, and granulocyte-macrophage colony-forming units). AA also inhibited growth of the bcr-abl-transformed cell line H7.bcr-abl A54. In contrast, a minimal effect of AA on inhibition of cell growth was observed in the parental nontransformed NSF/N1.H7 cell line. The antiproliferative effect of AA was associated with apoptosis. Gamma-linolenic acid, a precursor of AA, also inhibited cell growth, whereas other unsaturated and saturated fatty acids had no effect. Pharmacological inhibition of cyclooxygenase, lipooxygenase, and cytochrome P450 monooxygenase enzymes prior to exposure to AA did not rescue cells from the inhibitory effect of AA. Moreover, 5,8,11,14-eicosatetraynoic acid, a nonmetabolizable arachidonate analogue, also inhibited cell growth, suggesting that the effect of AA did not require further metabolism. Treatment with antioxidants prior to stimulation with AA was also ineffective in preventing its antiproliferative effect. Thus, AA inhibited proliferation of CML cells by inducing apoptotic cell death. The signaling mechanisms of AA-induced inhibition of cell growth appeared to be independent of its conversion into eicosanoids or free radical generation.

T-replete haploidentical allogeneic transplantation using post-transplantation cyclophosphamide in advanced AML and myelodysplastic syndromes.

Unmanipulated haploidentical transplantation (Haplo-SCT) using post-transplantation cyclophosphamide (PT-Cy) represents an alternative for patients with high-risk diseases lacking HLA-identical donor. Although it provides low incidences of GVHD, the efficacy of Haplo-SCT is still questioned, especially for patients with myeloid malignancies. Thus, we analyzed 60 consecutive patients with refractory (n=30) or high-risk CR (n=30) AML or myelodysplastic syndromes (MDSs) who underwent PT-Cy Haplo-SCT. The median age was 57 years (22-73 years), hematopoietic cell transplantation comorbidity index was ⩾3 in 38 patients (63%) and Haplo-SCT was the second allogeneic transplantation for 10 patients (17%). Although most of patients received PBSC as graft source (n=48, 80%), we found low incidences of grade 3-4 acute (2%) and severe chronic GVHD (4%). Among patients with high-risk CR diseases, 1-year non-relapse mortality, cumulative incidence of relapse, progression-free and overall survivals were 20%, 32%, 47% and 62%, respectively. In patients with refractory disease, corresponding results were 34%, 35%, 32% and 37%, respectively. We conclude that PT-Cy Haplo-SCT could provide promising anti-leukemic effect even in the setting of very advanced diseases. Thus, it represents a viable alternative for high-risk AML/MDS patients without HLA-identical donor.

Therapeutic relevance of CD34 cell dose in blood cell transplantation for cancer therapy.

PURPOSE: To review recent advances in peripheral-blood progenitor-cell (PBPC) transplantation in order to define the optimal cell dose required for autologous and allogeneic transplantation. MATERIALS AND METHODS: A search of MEDLINE was conducted to identify relevant publications. Their bibliographies were also used to identify further articles and abstracts for critical review. RESULTS: The CD34(+) cell content of a graft is regarded as an accurate predictor of engraftment success. Postchemotherapy autologous PBPC transplantation with >/= 5 x 10(6) CD34(+) cells/kg body weight leads to more rapid engraftment than does transplantation of lower cell doses. Further increases in transplant cell dose further accelerate platelet but not neutrophil engraftment. Evidence that long-term hematopoietic recovery may be more accurately predicted by the subpopulation of primitive progenitors transplanted suggests that the content of CD34(+)CD33(-) and long-term culture-initiating cells in cell collection samples may be important for predicting successful engraftment, particularly in patients with poor mobilization. Allogeneic transplantation has been limited by concerns regarding graft-versus-host disease and the use of hematopoietic growth factors in donors. The risk of graft rejection and engraftment failure after HLA-mismatched allogeneic transplantation may be overcome by intensive chemoradiotherapy and the infusion of large numbers of T cell-depleted hematopoietic stem cells. CONCLUSION: An optimal cell dose of >/= 8 x 10(6) CD34(+) cells/kg seems to be recommended for autologous PBPC transplantation. This dose facilitates the administration of scheduled chemotherapy on time and reduces the demand for other supportive therapies. A combination of growth factors may enable patients with poor mobilization to achieve a collection sufficient to allow transplantation. The optimum PBPC dose for allogeneic transplantation remains to be defined.

Autologous bone marrow transplantation in acute myelogenous leukemia.

Autologous bone marrow transplantation (ABMT) is a therapeutic approach that permits the administration of high-dose chemoradiotherapy followed by the infusion of the patient's own marrow, previously collected during remission and cryopreserved. In recent years, ABMT has been increasingly used as a treatment for acute leukemias. The mechanisms underlying leukemic relapse represent the most exciting and controversial aspects of ABMT. At least three factors may be responsible for leukemic relapse in patients receiving ABMT: (a) minimal residual disease; (b) leukemic cells reinfused with the graft; and (c) the lack of a graft-versus-leukemia effect. Techniques for pharmacological marrow decontamination, clinical results obtained with ABMT, and new perspectives opened by growth factors and cytokines are reviewed.

Biologic and phenotypic analysis of early hematopoietic progenitor cells in umbilical cord blood.

Umbilical cord blood (UCB) is an attractive potential alternative to bone marrow (BM) as a source of hematopoietic progenitor cells since the number of progenitors in UCB is similar or even greater than that in normal BM. It was the aim of the present study to analyze the degree of immaturity of UCB progenitor cells. UCB mononuclear (MNC) and/or CD34+ cells were tested for surface antigen phenotype, expression of cytokines receptor, effect of stem cell factor (SCF) on colony growth, resistance to mafosfamide and replating potential. We have found that 34.9 +/- 3.4% and 77.9 +/- 2.6% of UCB CD34+ cells did not express CD38 and CD45RA antigens, respectively, suggesting that UCB contains a high proportion of immature progenitor cells. By means of three-color analysis, the receptor for SCF was detected on the majority of the CD34+ HLA-DR+ subpopulation; in fact, 81.8% +/- 4.3% of CD34+ HLA-DR+ cells were defined as SCF(low) and 8.1 +/- 1.5% as SCF(high). Colony growth of MNC and CD34+ cells was enhanced by the addition of SCF to methylcellulose mixture, resulting in a statistically significant increase in CFU-GM and CFU-GEMM but not in BFU-E numbers. UCB progenitor cells showed a higher resistance to mafosfamide treatment, in comparison to BM; the addition of SCF to the culture medium resulted in a statistically significant increase in mafosfamide concentration required to inhibit 95% of colony growth (P < or = 0.05). Moreover, as shown by single colony transfer assays, the presence of SCF in primary cultures promoted a significantly higher replating potential for both untreated (42 +/- 3.3% vs 21 +/- 4.6%, P < or = 0.018) and mafosfamide-treated samples (62 +/- 5.6% vs 44 +/- 6.1%, P < or = 0.018). In conclusion, UCB is a source of progenitor cells with immature characteristics in terms of surface antigen expression, distribution of SCF receptor, resistance to mafosfamide and replating potential. Therefore, UCB progenitor cells represent an ideal candidate population for experimental programs involving gene transfer and ex vivo stem cell expansion.

Growth of human hematopoietic colonies from patients with myelodysplastic syndromes in response to recombinant human granulocyte-macrophage colony-stimulating factor.

The in vitro effect of recombinant human GM-CSF (rHuGM-CSF) was tested on bone marrow-derived multilineage (CFU-GEMM) as well as megakaryocytic (CFU-Mk), erythroid (BFU-E), and granulocyte-macrophage (CFU-GM) progenitors in a group (n = 16) of patients with myelodysplastic syndromes (MDS). Hematopoietic progenitor cell growth was markedly impaired in MDS patients as compared to normal controls (p less than 0.05, at least). Recombinant HuGM-CSF supported the growth of CFU-GEMM, CFU-Mk, and BFU-E at lower, equivalent, or slightly higher frequencies that those found in cultures plated with medium conditioned by peripheral blood leukocytes (PHA-LCM), but it was invariably ineffective in improving growth values. Recombinant HuGM-CSF supported the growth of granulocyte-macrophage colonies in 15 of 16 cases. The overall incidence (mean +/- SEM) of CFU-GM in cultures containing rHuGM-CSF (5 ng/ml) was significantly higher than the one found in cultures stimulated with PHA-LCM (40 +/- 15 vs. 17 +/- 7, p less than 0.05). Upon culture with rHuGM-CSF (5 ng/ml), in 5 of 15 patients de novo colony formation was observed (8 +/- 4) and in 4 of 15 patients CFU-GM growth (129 +/- 33) fell within normal range. Doses of rHuGM-CSF higher than 5 ng/ml did not result in a further increase of MDS-derived colony formation. It is concluded that rHuGM-CSF (a) does not improve the growth of CFU-GEMM, CFU-Mk, and BFU-E; (b) may completely restore the growth of CFU-GM in a subgroup of MDS patients; (c) while ineffective in improving anemia and thrombocytopenia, its in vivo in MDS may correct leukopenia through an effect at the level of granulocyte-macrophage progenitor cell compartment, at least in a subset of highly responsive patients.

Amifostine (WR-2721) selective protection against melphalan genotoxicity.

Amifostine (WR-2721) is an aminothiol compound dephosphorylated at the tissue site by alkaline phosphatase to the active metabolite, which is able to inactivate electrophilic substances and scavenge free radicals. Amifostine effects against melphalan-induced DNA strand breaks were studied in normal human white blood cells (WBC) and K562 leukemic cells using the single cell gel electrophoresis (SCGE) or Comet assay, a reported method for measuring DNA damage in individual cells. Prior to treatment (1 h, 37 degrees C) with increasing doses of melphalan, with or without S9, the cells were treated (15 min, 37 degrees C) with a control medium or amifostine (3 mg/ml). Treatment of normal and leukemic cells with melphalan induced a dose-dependent 'comet formation'. Melphalan-induced DNA damage follows a normal distribution in WBC. On the other hand, in K562, a significant proportion of undamaged cells remains even with doses at which mean DNA damage is serious. Pretreatment with WR-2721 protects WBC, but not K562, against the genotoxic effect of melphalan. Amifostine might even strengthen the action of the antiblastic drug against K562 cells. S9 addition appears to enhance melphalan effectiveness. SCGE appears as a suitable primary screening method for in vitro and in vivo studies on drug-DNA interactions and their modulations by endogenous/exogenous factors.

Growth of CD34+ acute myeloblastic leukemia colony-forming cells in response to recombinant hematopoietic growth factors.

In order to minimize the interactions of clonogenic cells with accessory cells and characterize the direct effect of recombinant hematopoietic growth factors (HGF) on acute myelogenous leukemia colony-forming cells (AML-CFU), the response of CD34+ AML-CFU to individual or combined recombinant HGF, i.e., interleukin-1 (IL-1), interleukin-3 (IL-3), interleukin-6 (IL-6), granulocyte-macrophage colony-stimulating factor (GM-CSF), granulocyte colony-stimulating factor (G-CSF), and macrophage colony-stimulating factor (M-CSF), was studied in 10 patients and compared with the growth response obtained from unfractionated marrow cells. IL-3 and GM-CSF had a similar stimulating activity on AML-CFU growth. G-CSF resulted the most efficient stimulus for colony formation and was additive or synergistic with IL-3 and GM-CSF, M-CSF, used alone, had a negligible stimulating activity. When CD34+ cells were used, IL-1 by itself had a low stimulating activity and displayed little or no synergy with IL-3, GM-CSF, and G-CSF. On the contrary, when unfractionated cells were used, IL-1 was very effective in inducing AML-CFU formation and was markedly synergistic with IL-3 and GM-CSF. These results show that IL-1-induced leukemic colony formation is prevalently mediated by accessory cells. IL-6 supported AML-CFU growth in seven of 10 cases, thus showing a direct effect on CD34+ leukemic cells, and enhanced the growth of IL-3-(+47 to +167%) and GM-CSF-dependent (+60 to +110%) AML-CFU. Recloning studies of single colonies demonstrated that primary CD34+ AML-CFU, stimulated by IL-3 and GM-CSF, generated secondary and tertiary colonies, whereas primary AML-CFU stimulated by G-CSF and IL-6 failed to give rise to secondary colonies, thus indicating a complete suppression of self-renewal. Sequential recloning of colonies grown in the presence of IL-3 + IL-6 demonstrated that addition of IL-6 and IL-3-containing plates resulted in a nearly complete suppression of self-renewal. In conclusion, these results demonstrate the heterogeneity of the CD34+ leukemic cell fraction and indicate the existence of complex regulatory events at the level of CD34+ leukemic cells. Data obtained from recloning experiments are of therapeutic interest in view of the clinical application of HGFs in the treatment of myeloid leukemias.