Tumor contamination of autologous peripheral blood stem/progenitor cell grafts occurs in a substantial proportion of high-risk breast cancer patients, and the possibility that such contamination may contribute to relapse has focused attention on methods for removal of the contaminating cells prior to transplantation. One such approach is positive selection of CD34+ cells. A fully automated immunomagnetic cell selection system has recently been introduced to facilitate the positive selection process. A multicenter randomized clinical trial was designed to evaluate the capacity of CD34+ cells isolated using the fully automated system to support prompt hematopoietic reconstitution following high-dose chemotherapy in high-risk breast cancer patients, as well as to assess the safety and tolerability of the CD34+ cell transplants. In recipients of isolated CD34+ cells, the median time to an absolute neutrophil count > or =500/microl was 10 days, a value identical to that observed in patients receiving unfractionated apheresis collections. In the isolated CD34+ cell recipients median time to a platelet count > or =20 000/microl was 12 days, compared with 10 days in the unfractionated cell group. There were no statistically significant differences between the groups in median time to neutrophil or platelet engraftment. Infusion of autologous cells was well tolerated by the study groups. There were no inter-group differences in the incidence of infections, need for platelet transfusions, or duration of hospitalization. Isolated CD34+ cells were high in purity and sufficient in number for use in autologous transplantation. The fully automated immunomagnetic cell selection system affords an efficient and time-saving option for isolation of CD34+cells to be used as autologous grafts in high-risk breast cancer patients, and the isolated CD34+ cells support undelayed hematopoietic reconstitution.

Between August 1994 and June 1999, 56 patients were prospectively randomized to receive ifosfamide 10 g/m2 + GM-CSF 5 microg/kg/day (IFO+GM-CSF n = 28) and cyclophosphamide 4 g/m2 + GM-CSF 5 microg/kg/day (CY+GM-CSF n = 28). Both groups were comparable for age, gender, diagnosis, disease stage and previous chemotherapy. The IFO+GM-CSF group demonstrated a shorter median interval between therapy and apheresis (10 days (8-14) vs 13 days (8-25) P = 0.002), median number of doses of GM-CSF (9 (7-13) vs 15 (9-31) P = 0.001), median of days with aplasia (0.5 (0-10) vs 6 (0-21) P = 0.001), median days with fever (0 (0-6) vs 3 (0-9) P = 0.006) and median of days using i.v. antibiotics (0 (0-11) vs 7.5 (0-19) P = 0.002). The median MNC yield was similar in both groups. The CD34+ cell yield was better in the CY+GM-CSF group (3.14 (0.9-11.8) vs 5.33 (0. 08-32)) but not at significant levels (P = 0.1). White blood cell hematopoietic recovery was more rapid in the CY+GM-CSF group (16 (10-22) vs 13 (10-24) P = 0.02). Platelet engraftment was similar in both groups. Costs of mobilization and transplantation were almost the same: $28 570 ($18 527-$47 028) and $30 020 ($17 281-$67 591), respectively (P = 0.9). There were no differences in disease-free survival and overall survival between both groups. Mild and transient non-hematological toxicity (hemorrhagic cystitis, decrease in serum creatinine clearance and CNS dysfunction) was seen most frequently in the IFO+GM-CSF group.

Sixty-one consecutive adult patients with leukaemia, primary myelofibrosis or myelodysplastic syndrome with an HLA-identical or one antigen mismatched family donor were randomised to allogeneic transplantation with PBPC or BM. Progenitor cells were mobilised into the blood by giving the donors 10 microg/kg/day G-CSF subcutaneously for 5-7 days. G-CSF was not given to patients after transplantation. The time to neutrophil counts >0.5 x 109/l was 17 days (95% CI 15.2-18.8 days) in the PBPC group compared to 23 (95% CI 20.3-25.7 days) in the BM group (P = 0.0005). The time to platelet counts >20 x 109/l was 13 days (95% CI 11.7-14.3 days) in the PBPC group and 21 days (95% CI 18.7-23.3 days) in the BM group (P = 0.0005). Acute GVHD of grades II-IV developed in six patients transplanted with PBPC and three patients transplanted with BM. The numbers of patients with chronic GVHD were 15 and 8, respectively. Transplant-related mortality and leukaemia-free survival showed no significant differences. Transplantation with PBPC appears preferable for the recipient due to faster neutrophil and platelet recovery. However, the final conclusion can not be drawn before long-term results on chronic GVHD and relapse incidence in longer randomised trials are available.

flt3-ligand (flt3-L) is a very effective mobilizer of hematopoietic stem cells (HSC) and is capable of inducing multilineage hematopoietic cell differentiation both in vivo and in vitro. We measured, by ELISA, the plasma peripheral blood flt3-L concentrations in 28 non-Hodgkin's lymphoma (NHL) patients before and after mobilization with three different mobilization regimens, including priming with cyclophosphamide (CY) plus G-CSF, CY plus GM-CSF, and CY plus GM-CSF (8 days) followed by G-CSF. We also determined the levels of flt3-L in the peripheral blood during four apheresis collections and 6 months after transplantation. The steady-state level of flt3-L in NHL patients (n = 18) who mobilized > or =2 x 10(6) CD34+ cells/kg in four apheresis collections was 34 +/- 4 pg/ml and was similar to the levels observed in 10 normal controls (27 +/- 7, p = 0.1) regardless of the mobilization protocol used. In contrast, patients who failed to mobilize a total of >0.4 x 10(6) CD34+ cells/kg in two consecutive apheresis collections (n = 10) had flt3-L levels of 106 +/- 11 pg/ml, significantly higher (p = 0.006) than that of the good mobilizers group, regardless of the mobilization protocol used. Similar results were observed in 29 multiple myeloma (MM) patients. A mean of 23.8 +/- 7.9 pg/ml and 450 +/- 85 pg/ml flt3-L was obtained in the good mobilizers (n = 24) and the nonmobilizers (n = 5) groups of patients, respectively. Statistical analysis revealed a significant difference (p = 0.0006) between the two groups of MM patients, but no correlation was observed between the levels of flt3-L and CD34+ cell/microl, in mobilized peripheral blood. Our results also suggest that measurement of plasma levels of flt3-L before mobilization can be clinically useful to predict for patients with poor mobilization outcome.

High-dose chemotherapy followed by autologous PBSC transplantation (PBSCT) has become an accepted form of therapy for a number of malignant hematologic diseases. The optimal method for the collection of PBSC is yet to be defined. Large-volume leukapheresis may be able to collect adequate numbers of PBSC with the patient undergoing fewer procedures. We routinely process 7 L of blood per leukapheresis. Hence, we elected to assess whether a modest increase in the blood volume processed would, on average, decrease the number of leukaphereses each patient needed to undergo to collect > or =2 x 10(6) CD34+ cells/kg body weight. Sixty patients were randomized to undergo 7 L leukaphereses (n = 31 patients; 87 leukaphereses) or 10 L leukaphereses (n = 29 patients; 81 leukaphereses). The median number of leukaphereses required per patient to collect the target number of CD34+ cells was two (range one to five) for both groups (p = 0.83). The median number of nucleated cells collected per patient was greater for the 10 L group (8.2 x 10(8)/kg versus 5.3 x 10(8)/kg, p = 0.005), as was the median number of mononuclear cells (MNC) (4.7 x 10(8)/kg versus 3.6 x 10(8)/kg, p = 0.0001), whereas there was no statistical difference between the groups for the median number of CD34+ cells collected per patient (3.2 x 10(6)/kg versus 3.7 x 10(6)/kg, p = 0.98). Therefore, over the 18-month period of this trial, the use of a 10 L leukapheresis volume did not decrease the number of leukaphereses performed compared with a 7 L leukapheresis volume.

Patient response to hematopoietic progenitor-cell mobilizing regimens seems to vary considerably, making comparison between regimens difficult. To eliminate this inter-patient variability, we designed a cross-over trial and prospectively compared the number of progenitors mobilized into blood after granulocyte-macrophage colony-stimulating factor (GM-CSF) days 1 to 12 plus granulocyte colony-stimulating factor (G-CSF) days 7 to 12 (regimen G) with the number of progenitors after cyclophosphamide plus G-CSF days 3 to 14 (regimen C) in the same patient.

Autologous transplantation with peripheral blood stem cells (PBSC) results in faster haematopoietic-cell repopulation than with bone marrow. We prospectively compared bone marrow and PBSC for allogeneic transplantation.

Immunofluorescence stainings for the CD10 antigen and terminal deoxynucleotidyl transferase (TdT) can be used for the detection of leukemic blasts in CD10+ precursor-B-acute lymphoblastic leukemia (precursor-B-ALL) patients, but can also provide insight into the regeneration of normal precursor-B-cells in bone marrow (BM). Over a period of 15 years, we studied the regeneration of CD10+, TdT+, and CD10+/TdT+ cells in BM of children with (CD10+) precursor-B-ALL during and after treatment according to three different treatment protocols of the Dutch Childhood Leukemia Study Group (DCLSG) which differed both in medication and time schedule. This study included a total of 634 BM samples from 46 patients who remained in continuous complete remission (CCR) after treatment according to DCLSG protocols VI (1984-1988; n = 8), VII (1988-1991; n = 10) and VIII (1991-1997; n = 28). After the cytomorphologically defined state of complete remission with CD10+ and CD10+/TdT+ frequencies generally below 1% of total BM cells, a 10-fold increase in precursor-B-cells was observed in protocol VII and protocol VIII, but not in protocol VI. At first sight this precursor-B-cell regeneration during treatment resembled the massive regeneration of the precursor-B-cell compartment after maintenance treatment, and appeared to be related to the post-induction or post-central nervous system (CNS) therapy stops in protocols VII and VIII. However, careful evaluation of the distribution between the 'more mature' (CD10+/TdT-) and the 'immature' (CD10+/TdT+) precursor-B-cells revealed major differences between the post-induction/post-re-induction precursor-B-cell regeneration (low 'mature/immature' ratio: generally <1.0), the post-CNS treatment regeneration (moderate 'mature/immature' ratio: 1.2-2.8), and the post-maintenance regeneration (high 'mature/ immature' ratio: 5.7-7.6). We conclude that a therapy stop of approximately 2 weeks is already sufficient to induce significant precursor-B-cell regeneration even from aplastic BM after induction treatment. Moreover, differences in precursor-B-cell regeneration patterns are related to the intensity of the preceding treatment block, with lower 'mature/immature' ratios after the highly intensive treatment blocks. This information is essential for a correct interpretation of flow cytometric immunophenotyping results of BM samples during follow-up of leukemia patients. Particularly in precursor-B-ALL patients, regeneration of normal precursor-B-cells should not be mistaken for a relapse.

The effects of granulocyte-colony stimulating factor (G-CSF) have been studied in several clinical settings. G-CSFs are widely used to stimulate the production of granulocytes and are well known to mobilize peripheral blood stem cells (PBSCs). However, very few studies have examined differences among G-CSFs. The aim of this study was to compare the mobilization of PBSCs induced by a standard dose of two G-CSFs following biweekly cyclophosphamide, doxorubicin, vincristine and prednisone (CHOP) therapy. Using a standard dose of G-CSF, we conducted a randomized, crossover trial that compared the efficacy of two kinds of G-CSF, glycosylated [lenograstim (2 micrograms/kg)] and mutated [nartograstim (1 microgram/kg)], on PBSC mobilization in 10 patients with non-Hodgkin's lymphoma after biweekly CHOP chemotherapy. Lenograstim (2 micrograms/kg) was more effective in shortening the duration of neutropenia than nartograstim (1 microgram/kg) (3.8 days vs. 5.0 days, p < 0.05, the number of days for the neutrophil count to reach 5 x 10(9)/l from nadir). The number of CD34+ cells and granulocyte-macrophage colony forming units (GM-CFU) was higher for lenograstim but no statistically significant difference between the two groups was found. Glycosylated G-CSF is more effective than mutated G-CSF in shortening the duration of neutropenia. As for the mobilization of CD34+ cells and the number of CFU-GM, there was a tendency to increase in the lenograstim group but no statistically significant differences were found.

We conducted a randomized trial in which we compared high-dose chemotherapy plus hematopoietic stem-cell rescue with a prolonged course of monthly conventional-dose chemotherapy in women with metastatic breast cancer.

In two separate lymphoma populations, we examined immune reconstitution following high dose chemotherapy (HDT) and bone marrow transplantation (BMT). In the first study we followed immune reconstitution for one year after HDT and BMT. In the second study we examined the ability of the orally active immunomodulator, Bestatin to augment immune reconstitution following HDT and BMT. The studies on immune reconstitution following HDT and BMT were undertaken in a cohort of non-Hodgkin's lymphoma (NHL) patients (n = 35) and examined the peripheral blood (PB) leukocyte subsets and their in vitro functions. Our results demonstrate that monocyte and natural killer (NK) cell engraftment occurred more rapidly then did T cell reconstitution. We also observed a significant decrease in the CD4:CD8 ratio post-transplantation as compared to normal PB donors due to a decrease in CD4+ cells. In addition, following HDT and BMT, measures of T cell function (phytohemagglutinin [PHA] mitogenesis) and T helper cell activity (pokeweed mitogen [PWM] mitogenesis) were consistently depressed as compared to cells from normal PB. Further, we demonstrate a correlation between the loss of T cell function and the frequency of circulating monocytes, suggesting a cause-effect relationship. Despite the dysfunction in T cells following HDT and BMT, immune-modulating agents can still augment the immune function. One such drug is Bestatin (ubenimex), an inhibitor of aminopeptidase (AP) that binds to CD13 on macrophage/monocytes. To examine its immune modulatory activity after HDT and BMT, a dose finding (10, 30, 90 and 180 mg/day) phase Ib trial was conducted with 30 Hodgkin's disease (HD) and NHL patients who received no drug (control), or Bestatin daily for 60 days following BMT. In these studies, Bestatin administration was initiated when the absolute neutrophil count was greater than 250/mm3 on two consecutive days. These studies revealed that Bestatin significantly increased the PHA and PWM responses in a dose-dependent manner. Flow cytometric analysis revealed a significant increase in NK cells (CD56+), B cells (CD19+), as well as the CD4:CD8 cell ratio. The latter observation was associated largely with a depression in the percent of CD8+ T cells as opposed to an increase in CD4+ T cells. We conclude that despite the peripheral tolerance observed following HDT and BMT, Bestatin could significantly increase some, but not all, immune surrogates.

Mobilization of peripheral blood cell progenitor cells was investigated in 36 healthy sibling donors using three different split doses of glycosylated rhG-CSF (lenograstim). The donors were randomized into three groups: group 1 was given lenograstim at 8, group 2 at 11 and group 3 at 15 micrograms/kg/day in two split doses, subcutaneously for 4 and 5 days, respectively. Leukapheresis was performed on day 4 or 5 depending on the WBC and CD34+ cell count. We were able to demonstrate that there was a significant correlation between circulating CD34+ cells on the day of harvest and CD34+ cells in the apheresis products in all three groups. The number of CD34+ cells pre-apheresis was inversely correlated with age in group 1 and group 2. However, in group 3, the number of CD34+ cells pre-apheresis did not correlate with age. There was also a difference between the number of progenitor cells mobilized in the three dose groups regarding the time of harvest. Apheresis was performed in groups 1 and 2 on day 5 of mobilization in order to obtain a sufficient number of stem cells for allogeneic transplantation. In contrast, with the split dose of 15 micrograms/kg/day, harvest could be routinely performed on day 4 of stimulation. We conclude that lenograstim given twice a day at doses of 8, 11 and 15 micrograms/kg/day provided different CD34+ cell yields in normal donors, in particular, with regard to the time of harvest. The number of CD34+ cells pre-apheresis was not correlated with age in the group of donors mobilized with a split dose of 15 micrograms/kg/day, indicating that this dosage might also be suitable for older donors.

The purpose of this study was to develop a cost-effective protocol for the mobilization of peripheral blood stem cells (PBSC) in patients with malignancy. Thirty consecutive patients were randomized to mobilize PBSC with the late addition of a standard 250 microg dose of G-CSF (Neutrogen) from day 8 or early addition of the same dose of G-CSF from day 2, following cyclophosphamide (CY) 4 g/m2. The median yield of CD34+ cells from evaluated patients was 7.87 x 10(6)/kg (range, 2.06-27.25), collected in a median of four apheresis (range, 2-9). Target CD34 + cell doses > or = 2.0 x 10(6)/kg were achieved in all patients able to be evaluated. There were no statistically significant differences in CD34+ cell yields or toxicities. Overall engraftment occurred with median days to neutrophils > or = 0.5 x 10(9)/L or platelets > 20 x 10(9)/L of 11 and 17 days, respectively. However, the duration of G-CSF administration was markedly shorter in the late use of G-CSF group than in the early use of G-CSF group, with a median of 9 days compared with 15 days (p<0.001). PBSC harvesting after priming with CY plus delayed use of G-CSF made it a safe and cost-effective procedure.

The purpose of this study was to investigate whether storing mobilized peripheral blood progenitor cell (PBPC) collections overnight before CD34+ selection may delay platelet count recovery after high-dose chemotherapy and CD34+-enriched PBPC re-infusion. Lymphoma patients underwent PBPC mobilization with cyclophosphamide 4 g/m2 i.v. and G-CSF 10 microg/kg/day subcutaneously. Patients were prospectively randomized to have each PBPC collection enriched for CD34+ cells with the CellPro CEPRATE SC System either immediately or after overnight storage at 4 degrees C. Thirty-four patients were randomized to overnight storage and 34 to immediate processing of PBPC; 15 were excluded from analysis due to tumor progression or inadequate CD34+ cell mobilization. PBPC from 23 patients were stored overnight, while 30 subjects underwent immediate CD34+ selection and cryopreservation. Median yield of CD34+ enrichment was 43.6% in the immediate processing group compared to 39.1% in the overnight storage group (P = 0.339). Neutrophil recovery >500 x 10(9)/l occurred a median of 11 days (range 9-16 days) in the overnight storage group compared to 10.5 days (range 9-21 days) in the immediate processing group (P = 0.421). Median day to platelet transfusion independence was 13 (range 7-43) days in the overnight storage group vs 13.5 (range 8-35) days in those assigned to immediate processing (P = 0.933). We conclude that storage of PBPC overnight at 4 degrees C allows pooling of consecutive-day collections resulting in decreased costs and processing time without compromising neutrophil and platelet engraftment after infusion of CD34+-selected progenitor cells. Bone Marrow Transplantation(2000) 25, 559-566.

Forty healthy adult donors underwent marrow (BM) as well as peripheral blood (PBSC) stem cell collections for their HLA-identical adult siblings with hematologic malignancies. BM was harvested on day 1 (target 3 x 108 nucleated cells/kg, 10 microg/kg lenograstim (glycosylated G-CSF) administered on days 2-6, and a single leukapheresis performed on day 6. The blood volume processed was the higher of 200% donor blood volume or 10 liters. The total nucleated cell (TNC) yields from PBSC were 1.1- to 4.3-fold higher than BM (median 7.0 vs 3.1 x 10(8)/kg, P < 0.0001). Although BM contained a higher proportion of CD34+cells (1.3% vs 0.7%, P < 0. 0001) and a comparable proportion of CD3+ cells (median 29% vs 26%, P = 0.4), the absolute numbers of CD34+ and CD3+ cells and their subsets were several times higher in PBSC. There was a poor correlation between BM and PBSC CD34 and TNC numbers, but a significant correlation between BM and PBSC CD3 numbers. Only five of 40 BM harvests contained >/=2 x 10(6) CD34+ cells/kg compared with 35 of 40 PBSC harvests (P < 0.0001). We conclude that the numbers of progenitor and immunocompetent cells in PBSC are several times higher than in BM. It is possible to collect adequate numbers of progenitor cells from blood after lenograstim stimulation more frequently than from marrow, and donors yielding low quantities of progenitor cells from BM usually deliver better quantities from PBSC. Bone Marrow Transplantation (2000) 25, 501-505.

Human herpesvirus 8 (HHV-8), also known as Kaposi's sarcoma-associated herpesvirus (KSHV), has recently been identified within the bone marrow dendritic cells of multiple myeloma (MM) patients. This virus contains homologues to human cytokines such as IL-6 that could potentially stimulate myeloma cell growth and contribute to disease pathogenesis. Since mobilization chemotherapy may increase circulating dendritic cell numbers, we searched for HHV-8 in peripheral blood mononuclear cells (PBMCs) before and after mobilization chemotherapy given to MM patients. Furthermore, we determined if autograft purging using the CEPRATE SC device would reduce the percentage of HHV-8 infected stem cell products. Only two of the 39 PBMC samples collected prior to mobilization chemotherapy contained PCR detectable virus, yet nine of 37 PBMCs collected on the first day of leukapheresis had detectable HHV-8 (P = 0.016). HHV-8 was more frequently identified in autograft products before vs after Ceprate SC selection (40% vs 15%, P = 0.016). Although the role HHV-8 plays in myeloma pathogenesis remains unclear, these results imply that mobilization chemotherapy increases the numbers of circulating HHV-8-infected dendritic cells within the peripheral blood. In addition, CD34 selection of autograft products in MM patients may reduce the reintroduction of virally infected cells following high-dose chemotherapy. Bone Marrow Transplantation (2000) 25, 153-160.

To compare hematologic recovery in patients receiving allogeneic blood cell transplantation (BCT) with those receiving allogeneic bone marrow transplantation (BMT).

CD(34+)-enriched peripheral blood stem cells (PBSC) are increasingly being used as an autograft in patients with multiple myeloma (MM). The rationale for the use of the CD34+-enriched fraction in MM is the ability to obtain a graft with a significant reduction of contamination by plasma cells. However, the effect of such a manipulation on the proliferating potential of the engrafted cells is not known. We wished to study, as part of a randomized trial comparing the outcome in MM patients transplanted with either CD(34+)-enriched cells or unfractionated PBSC, the primitive hematopoietic cell content of the autografts using long-term culture initiating cell (LTC-IC) assays in 7 MM patients. In 3 patients CD(34+)cell-enriched fraction was compared to unfractionated PBSC whereas in the remaining 4 patients the LTC-IC assay was performed on total PBSC. The mean percentage of CD34+ cells of the CD34+ selected fraction in three patients was 82% (range 71%-96%) whereas the same percentage in PBSC varied from 0.6% to 10% in 4 patients (mean: 4.2%). Out of three patients transplanted with CD34+ cell fraction, two patients were found to have a very similar LTC-IC generating potential in their CD34+ versus PBSC fractions as this was assessed by the clonogenic cell output at week+5 per 10(4) CD34+ cells initiating the culture (PBSC: 92 and 168 and CD34+ fraction: 102 and 16, respectively) whereas one patient had a slightly different values (PBSC: 51 and CD34+ fraction: 103). When the PBSC fraction was compared in all 7 patients, the LTC-IC generation potential was very heterogenous, varying from 1.4 to 168. To determine if the selection procedure influences the numbers of LTC-IC's in both fractions, we have performed limiting dilution assays to determine both the frequency of distribution of hematopoietic colonies and the frequency of LTC-IC's in two patients. The frequency of distribution of hematopoietic colonies was linear in both CD34+ and PBSC fractions as was the frequency of LTC-IC when the corrections were made with regard to the CD34+ cell-content of the cultures (1/20). Our results indicate that the CD34+ selection procedure used in all three patients (Ceprate) is not deleterious for the generation of LTC-IC's and these findings support the rationale for the use of this procedure in multiple for the purposes of tumor depletion.

Thrombocytopoiesis of 21 multiple myeloma patients undergoing single or double transplant regimen was characterized by measuring the level of reticulated platelets and plasma glycocalicin. Since reticulated platelets are an index of thrombopoietic activity and glycocalicin plasma values are related to platelet damage and turnover, it may be possible to perform a novel type of analysis of the thrombopoietic compartment during the mobilizing regimen and during transplant-related chemotherapy. Patients underwent mobilizing therapy and first transplant. Some randomized patients also underwent a second transplant with mobilized peripheral blood stem cells. The results show that the percentage of reticulated platelets decreased after therapy and then gradually increased in the recovery phase either during first or second transplant. By contrast, the percentage of reticulated platelets increased until day +8 and then gradually decreased during the mobilizing regimen. The glycocalicin index (glycocalicin plasma value normalized for the individual platelet count) increased significantly both during the course of mobilization and after transplant-related chemotherapy when the platelet number was at its nadir. However, the glycocalicin index was more elevated after transplant-related chemotherapy than after the mobilizing regimen. Our findings suggest that chemotherapy-related thrombocytopenia may be due to a dual mechanism: thrombocytopenia results from decreased platelet production in addition to increased platelet damage and possible destruction.

Several agents, used either alone or in combination, have been shown to boost absolute numbers of PMLs and CD34+ stem cells in PB. In this study, we compared the effects of three different treatments, G-CSF, HDMP, and G-CSF + HDMP, for neutropenic patients (absolute PML count < 0.5 x 10(9)/l) who were on maintenance therapy for ALL with a control group who received no treatment. Hematological parameters, PB-CD34+33- and -CD34+ HLA-DR- stem cell numbers, and serum IL-3 levels were measured prior to, and a week after, the first day of treatment. WBC and absolute PML counts were significantly increased compared to pretreatment values in all treatment groups. However, peripheral CD34+ 33- and CD34+ HLA-DR stem cell numbers and serum IL-3 levels were increased significantly only in the G-CSF group. There was also a significant increase in serun IL-3 levels in the G-CSF + HDMP group. This study suggests that G-CSF may induce an increase in peripheral stem cell numbers, and that it supports hemopoietic recovery by increasing IL-3 in patients with chemotherapy-induced neutropenia.