Flow cytometric enumeration of CD34+ hematopoietic stem and progenitor cells. European Working Group on Clinical Cell Analysis

The need for a rapid and reliable marker for the engraftment potential of hematopoietic stem and progenitor cell (HPC) transplants has led to the development of flow cytometric assays to quantitate such cells on the basis of their expression of CD34. The variability associated with enumeration of low-frequency cells (i.e., as low as 0.1% or 5 cells/microl) is exceedingly large, but recent developments have improved the accuracy and precision of the assay. Here, we review and compare the major techniques. Based on the current state of the art, we recommend 1) bright fluorochrome conjugates of class II or III monoclonal antibodies (mAbs) that detect all glycoforms of CD34, 2) use of a vital nucleic acid dye to exclude platelets, unlysed red cells, and debris or use of 7-amino actinomycin D to exclude dead cells during data acquisition, 3) counterstaining with CD45 mAb to be included in the definition of HPC, 4) during list mode data analysis, Boolean gating to resolve the CD34+ HPCs from irrelevant cell populations on the basis of the low levels of CD45 expression and low sideward light-scatter signals of HPCs, 5) inclusion of CD34dim and CD34bright populations in the CD34+ cell count, 6) omission of the negative control staining, and 7) for apheresis products, enumeration of at least 100 CD34+ cells to ensure a 10% precision. Unresolved technical questions are 1) the replacement of conventional dual-platform by single-platform assay formats, i.e., derivation of absolute CD34+ cell counts from a single flow cytometric assessment instead of from combined flow cytometer (percent CD34+) and hematology analyzer (absolute leukocyte count) data, 2) the cross-calibration of the available single-platform assays, and 3) the optimal method for sample preparation. An important clinical question to be addressed is the definition of the precise phenotypes and required numbers of HPCs responsible for short- and long-term recovery to optimize HPC transplant strategies.

Mobilization of peripheral blood progenitor cells by disease-specific chemotherapy in patients with soft tissue sarcoma

We investigated peripheral blood progenitor cell (PBPC) mobilization by disease-specific chemotherapy in patients with metastatic soft tissue sarcoma (STS). Nine patients, five females and four males, aged 12-51 years, pretreated by one to nine courses of cytotoxic chemotherapy, underwent STS-specific mobilization followed by G-CSF at 5 microg/kg/day. PBPC were collected by 19 conventional-volume aphereses (8-12 l) with one to four procedures in individual patients. Leukaphereses started on median day 15 (range 13-18) from the first day of mobilization chemotherapy at medians of 25.8 x 10(3) WBC/microl (6.8-46.9), 3.5 x 10(3) MNC/microl (1.1-8.8), 122 x 10(3) platelets/microl (72-293) and 30.7 CD34+ cells/microl (6.7-207.8). Cumulative harvests resulted in medians of 4.6 x 10(8) MNC/kg (3.0-6.4), 2.9 x 10(6) CD34+ cells/kg (1.1-11.1) and 12.0 x 10(4) CFU-GM/kg (2.0-37.8). Eight patients underwent high-dose chemotherapy (HDCT) followed by PBPC rescue. Seven patients recovered hematopoiesis at medians of 12 days (8-15) for ANC >0.5 x 10(3)/microl and 14 days (8-27) for platelets >20 x 10(3)/microl. One patient, who received 1.6 x 10(6) CD34+ cells/kg, exhibited delayed ANC recovery on day +37 and failed to recover platelets until hospital discharge on day +55. We conclude that in patients with metastatic STS, who are pretreated by standard chemotherapy, PBPC can be mobilized by a further course of STS-specific chemotherapy plus G-CSF. One to four conventional-volume aphereses result in PBPC autografts that can serve as hematopoietic rescue for patients scheduled for HDCT.

[Flow cytometry quality control in stem cell separation]

Peripheral blood-derived haematopoietic stem cells (PBSC) are used as an alternative to bone marrow stem cells for autologous transplantation. One of the most important prerequisites for successful PBSC separation is the precise determination of the optimal separation days. As we previously demonstrated that neither PLT counts nor WBC counts in the peripheral blood (PB) are of predictive value for the amount of colony-forming cells (CFC) in the patients' PB, we started a daily monitoring of CD-34-positive cells to determine the beginning of the separation series. In addition to routine cell counts and CFU testing we assessed the number of CD 34+ in the PBSC concentrates to ascertain the number of separations needed for each patient. Due to the strong correlation of CD 34+ cells to CFC it is possible to predict the number of CFC collected within 2 h after finishing the PBSC separation and to calculate the efficiency of the separation for quality control.

Circulating CD34-expressing cells: German Proficiency Testing Survey

Determination of absolute numbers of CD34-expressing cells is critical in the setting of peripheral blood stem and progenitor cell transplantation/reinfusion. The diagnostic value of the parameter, CD34-expressing cells/microliter, has been validated. A survey of CD34-expressing cells has been integrated into a series of flow cytometry proficiency testing surveys (reticulocytes, lymphocytes, leukemia, and lymphoma) that we have established in Germany. Commercially available, modified, stabilized myeloblastic leukemia cells (KG1a cell line) spiked at different numbers into two normal blood samples were sent out, and report forms were returned from 50 of 58 participants. With a predicted percentage of CD34-expressing cells of 0.5% (sample A) and of 0.25% (sample B), the respective mean values analyzing data from 44 participants returning the completed forms were 0.49% (sample A) and 0.29% (sample B). The coefficients of variation were 57% and 83%, respectively. Engineered samples based on normal blood and on commercially available stabilized modified KG1a cells seem to be reliable material for external quality assessment surveys of CD34-expressing cells.

De novo malignancies after liver transplantation using tacrolimus-based protocols or cyclosporine-based quadruple immunosuppression with an interleukin-2 receptor antibody or antithymocyte globulin

BACKGROUND

Although conventional immunosuppression after liver transplantation consists of cyclosporine A (CsA), steroids, and azathioprine, recently introduced protocols entail CsA-based quadruple induction protocols or tacrolimus-based combinations. These protocols aim to reduce the rejection rate and the considerable morbidity related to the side effects of additional immunosuppressive treatment, but have not yet been analyzed regarding their long term de novo neoplastic risk.

METHODS

From September 1988 to May 1994, 500 liver transplantations were performed in 458 patients. The median follow-up was 50 months (range, 0.3-97 months) for all patients. Conventional triple therapy was implemented in 25 patients, CsA-based quadruple induction therapy using an antilymphocyte globulin preparation (ATG) in 190 patients, an interleukin-2 receptor antibody (BT563) in 141 patients, and tacrolimus-based dual or triple immunosuppression in 102 patients. The different protocols were evaluated in four randomized and two nonrandomized prospective trials.

RESULTS

De novo neoplasias were detected in 33 patients (7.2%) and were comprised of lymphomas (n = 7), skin malignancies (n = 8 lesions in 7 patients), intraepithelial neoplasias of the cervix uteri (n = 7), breast carcinoma (n = 3), lung carcinoma (n = 3), and other malignancies (n = 6). The incidence of de novo neoplasias did not differ in the different trial arms. Only a positive T-crossmatch and a low CD4+/CD8+ ratio in patients receiving CsA-based immunosuppression demonstrated a significant correlation with the development of a de novo tumor in a multivariant logistic regression analysis.

CONCLUSIONS

The development of de novo neoplastic diseases after liver transplantation with the use of CsA-based quadruple induction protocols or tacrolimus-based regimens for immunosuppresion was assessed over the long term. Recently introduced immunosuppressive protocols did not alter the posttransplant de novo tumor rate. Patients with a low CD4+/CD8+ ratio during CsA-based therapy or a positive T-crossmatch were identified to be at an increased risk for the development of a de novo malignancy.

De novo malignancies after liver transplantation using tacrolimus-based protocols or cyclosporine-based quadruple immunosuppression with an interleukin-2 receptor antibody or antithymocyte globulin

BACKGROUND

Although conventional immunosuppression after liver transplantation consists of cyclosporine A (CsA), steroids, and azathioprine, recently introduced protocols entail CsA-based quadruple induction protocols or tacrolimus-based combinations. These protocols aim to reduce the rejection rate and the considerable morbidity related to the side effects of additional immunosuppressive treatment, but have not yet been analyzed regarding their long term de novo neoplastic risk.

METHODS

From September 1988 to May 1994, 500 liver transplantations were performed in 458 patients. The median follow-up was 50 months (range, 0.3-97 months) for all patients. Conventional triple therapy was implemented in 25 patients, CsA-based quadruple induction therapy using an antilymphocyte globulin preparation (ATG) in 190 patients, an interleukin-2 receptor antibody (BT563) in 141 patients, and tacrolimus-based dual or triple immunosuppression in 102 patients. The different protocols were evaluated in four randomized and two nonrandomized prospective trials.

RESULTS

De novo neoplasias were detected in 33 patients (7.2%) and were comprised of lymphomas (n = 7), skin malignancies (n = 8 lesions in 7 patients), intraepithelial neoplasias of the cervix uteri (n = 7), breast carcinoma (n = 3), lung carcinoma (n = 3), and other malignancies (n = 6). The incidence of de novo neoplasias did not differ in the different trial arms. Only a positive T-crossmatch and a low CD4+/CD8+ ratio in patients receiving CsA-based immunosuppression demonstrated a significant correlation with the development of a de novo tumor in a multivariant logistic regression analysis.

CONCLUSIONS

The development of de novo neoplastic diseases after liver transplantation with the use of CsA-based quadruple induction protocols or tacrolimus-based regimens for immunosuppresion was assessed over the long term. Recently introduced immunosuppressive protocols did not alter the posttransplant de novo tumor rate. Patients with a low CD4+/CD8+ ratio during CsA-based therapy or a positive T-crossmatch were identified to be at an increased risk for the development of a de novo malignancy.

De novo malignancies after liver transplantation using tacrolimus-based protocols or cyclosporine-based quadruple immunosuppression with an interleukin-2 receptor antibody or antithymocyte globulin

BACKGROUND

Although conventional immunosuppression after liver transplantation consists of cyclosporine A (CsA), steroids, and azathioprine, recently introduced protocols entail CsA-based quadruple induction protocols or tacrolimus-based combinations. These protocols aim to reduce the rejection rate and the considerable morbidity related to the side effects of additional immunosuppressive treatment, but have not yet been analyzed regarding their long term de novo neoplastic risk.

METHODS

From September 1988 to May 1994, 500 liver transplantations were performed in 458 patients. The median follow-up was 50 months (range, 0.3-97 months) for all patients. Conventional triple therapy was implemented in 25 patients, CsA-based quadruple induction therapy using an antilymphocyte globulin preparation (ATG) in 190 patients, an interleukin-2 receptor antibody (BT563) in 141 patients, and tacrolimus-based dual or triple immunosuppression in 102 patients. The different protocols were evaluated in four randomized and two nonrandomized prospective trials.

RESULTS

De novo neoplasias were detected in 33 patients (7.2%) and were comprised of lymphomas (n = 7), skin malignancies (n = 8 lesions in 7 patients), intraepithelial neoplasias of the cervix uteri (n = 7), breast carcinoma (n = 3), lung carcinoma (n = 3), and other malignancies (n = 6). The incidence of de novo neoplasias did not differ in the different trial arms. Only a positive T-crossmatch and a low CD4+/CD8+ ratio in patients receiving CsA-based immunosuppression demonstrated a significant correlation with the development of a de novo tumor in a multivariant logistic regression analysis.

CONCLUSIONS

The development of de novo neoplastic diseases after liver transplantation with the use of CsA-based quadruple induction protocols or tacrolimus-based regimens for immunosuppresion was assessed over the long term. Recently introduced immunosuppressive protocols did not alter the posttransplant de novo tumor rate. Patients with a low CD4+/CD8+ ratio during CsA-based therapy or a positive T-crossmatch were identified to be at an increased risk for the development of a de novo malignancy.

Imprecision of counting CFU-GM colonies and CD34-expressing cells

Determinations of committed haemopoietic progenitor cells, namely CFU-GM (colony-forming unit-granulocyte/macrophage) and of CD34-expression haemopoietic cells as assessed by multiparameter flow cytometry are routine diagnostic tools in haemopoietic cell therapy. Generally, the tests are used to optimise the timing and management of cytapheresis and to assess the engraftment potential of the harvested cells. Both measurements, however, are at best surrogate markers, as an adequate routine test which effectively assesses the short- and long-term repopulating haemopoietic cell is not available. Nonetheless, cell threshold doses have been established. Above these thresholds rapid engraftment is almost invariable but below these thresholds the outcome is variable. In this study we have focussed on the imprecision in counting haemopoietic cells, as assessed as CFU-GM and as CD34-expressing cells. The data on both tests have been analysed from six European institutions. The coefficient of variation in CFU-GM colony counting was about 30%, whereas the coefficient of variation in flow cytometric counting of CD34-expressing cells was about 10%. These data suggest that the technical imprecision in enumerating progenitor cells, particularly CFU-GM, at low levels, might make a major contribution to the clinical variability observed after transplantation of sub-threshold progenitor cell dose.

Two subsets of peripheral blood plasma cells defined by differential expression of CD45 antigen

The purpose of this study was to optimize the flow cytometric determination of circulating normal and malignant plasma cells (PC). We investigated peripheral blood (PB) samples of 65 patients with multiple myeloma or monoclonal gammopathy of unknown significance and 47 control subjects using CD38, CD45, B-B4, CD56, VLA-4, VLA-5 and CD19 monoclonal antibodies (MoAbs). Mono- or polyclonality was determined by staining of intracellular kappa and lambda light chains. Two subpopulations of PBPC were distinguished by differential expression of CD45. CD45 positive (CD45+) PC showed a more immature morphology and were detected in all groups. They were polyclonal in the control subjects and either poly- or monoclonal in the myeloma patients. In contrast, CD45 negative (CD45-) PBPC only occurred in myeloma patients and were consistently monoclonal, their presence being significantly associated with high disease activity (P < 0.001). Although detection of CD45- PBPC using CD38 or B-B4 MoAbs lead to similar results. CD45+ PBPC often were recognized to a lesser extent by B-B4 than by CD38 MoAbs. In conclusion, normal and malignant circulating PC can reliably be identified using CD38 and CD45 MoAbs. CD45 expression separates PBPC into two subsets of which the CD45- one only occurs in myeloma patients.

Recent progress on the role of Axl, a receptor tyrosine kinase, in malignant transformation of myeloid leukemias

Receptor tyrosine kinases (RTK) play an important role in the signal transduction of normal and malignant cells. There are different families of RTKs which are mainly characterized by differences in the ligang-binding extracellular domains. Axl (or UFO/Ark) is the first member of a new class of RTK with two fibronectin type III domains and two immunoglobulin-like domains present at the extracellular domain. The axl-gene has been isolated by means of gene transfection studies using DNA of patients with chronic myelogeneous leukemia. For a previous and the present study, we used a sensitive reverse-transcriptase polymerase chain reaction assay to detect axl's mRNA in cells from normal and malignant hematopoietic tissue. Axl's mRNA expression was mainly detected in myelo-monocytic cells, whereas much weaker transcription was seen in lymphatic cells and in lymphatic leukemias. In normal bone marrow, axl was heavily transcribed in marrow stromal cells. Further, we analysed Axl protein expression using monoclonal antibody M50 in peripheral stem cell harvests; in most harvests, no co-expression of CD34 and Axl was detected. However, in one patient with AML in complete remission, Axl was co-expressed on 80% of the CD34-positive population. These data show that axl is preferentially expressed in monocytes and stromal cells. Furthermore, a fraction of CD34-positive progenitor cells may express Axl. The exact mechanism for transformation of myeloid progenitor cells through Axl, however, remains to be determined.

Peripheral blood progenitor cell collection during hematopoietic recovery following autologous blood progenitor cell transplantation

BACKGROUND AND OBJECTIVES

Peripheral blood progenitor cells (PBPC) are increasingly used for autologous transplantation after high-dose radio/chemotherapy in patients suffering from cancer. PBPC are usually collected after mobilization with conventional-dose chemotherapy plus growth factor. However, it is conceivable to perform leukapheresis for the second autograft during recovery of hematopoiesis after the first course of HDCT/ABPCT.

MATERIALS AND METHODS

We treated two patients this way. In the first, with germ cell cancer, six 12-liter leukaphereses yielded 1.8 x 10(6) CD34+ cells/kg after mobilization with cis-platinum, etoposide and ifosfamide (PEI) plus granulocyte colony-stimulating factor (G-CSF). The second patient, with relapsed Hodgkin's disease, underwent PBPC collection after treatment with dexamethasone, carmustine, etoposide, cytarabine and melphalan (DexaBEAM) plus G-CSF. Due to excellent mobilization, 8.5 x 10(6) CD34+ cells/kg were collected by one 12-liter leukapheresis. Both patients then underwent PBPC collection during hematopoietic recovery following HDCT and ABPCT.

RESULTS

In patient 1, following HDCT and ABPCT, three 12-liter aphereses resulted in 0.7 x 10(6) CD34+ cells/kg. In patient 2, also after HDCT and ABPCT, a second autograft with 3.2 x 10(6) CD34+ cells/kg was harvested by a single 10-liter apheresis. No adverse effects were seen in either patient during apheresis following ABPCT. To our knowledge this is the first report dealing with PBCT collection during hematopoietic recovery following HDCT and ABPCT.

CONCLUSIONS

(1) PBPC harvesting is feasible and well tolerated in this setting. (2) In appropriate patients with efficient PBPC mobilization after conventional-dose chemotherapy, a further PBPC autograft can be collected during recovery of hematopoiesis after ABPCT, serving as a rescue for a second course of HDCT.

In-vitro clonogenicity of mobilized peripheral blood CD34-expressing cells: inverse correlation to both relative and absolute numbers of CD34-expressing cells

The determination of CD34-expressing cells by multiparameter flow cytometry is now widely used to estimate the reconstitution potential of cells harvested by cytapheresis for peripheral blood stem cell and progenitor cell transplantation. There is a correlation between the number of CD34-expressing cells collected and committed progenitor cells (CFU-GM and BFU-E) capable of forming colonies in vitro, but there is considerable variation in the proportion of CD34-expressing cells capable of clonogenic growth. The data in this study of 782 cytapheresis samples indicates that there is a negative correlation between the clonogenicity of the CD34-expressing cells and the absolute number or the proportion of CD34-expressing cells within the harvest. In 116 samples the proportion of CD34-expressing cells co-expressing the CD45-RA-antigen (a subset of CD34-expressing cells which includes virtually all clonogenic cells in terms of CFU-GM) was determined, but this did not help to identify the clonogenicity of a given sample. These findings may have clinical relevance, particularly when mobilization is judged to be relatively poor or when a good harvest is to be divided for multiple high-dose procedures.

Expression of class I, II and III epitopes of the CD34 antigen by normal and leukemic hemopoietic cells

Peripheral blood samples from 115 consecutive patients and bone marrow samples from 9 healthy donors were studied for percentages of CD34-expressing cells, quantitative expression of various CD34 epitopes as defined by fluorescence mean channel, and mutual inhibition of the different CD34 monoclonal antibodies detecting the various CD34 epitopes. The study focused only on samples from patients with presumably elevated numbers of CD34-expressing cells, due to the nature of the disease. Samples from patients with chronic myeloproliferative syndromes, acute leukemias from nonhematological cancer patients during mobilization with filgrastim, and normal bone marrow samples were studied. Elevated numbers of CD34-expressing cells (> 0.04% of all nucleated cells) were detected in 111 of 124 patients. An almost identical expression of CD34 epitopes as detected by phycoerythrin-conjugated monoclonal antibodies QBEND-10, 8G12, and ICH3 were detected, whereas expression of the IMMU409 epitope was detected in only a few samples (19 of 111). Reactivity of IMMU-133 was almost identical to that of QBEND-10. Reactivity of BIRMA-K3, the only CD34 monoclonal used in this study, but not described in previous workshops, was almost identical to that of 8G12. From studies on mutual inhibition of binding, two families of CD34 epitopes are defined. The first family is comprised of QBEND-10, IMMU-133, My-10, and ICH-3, and the second one is comprised of only 8G12 and BIRMA-K3.

Interference of blood leucocytes in the measurements of immature red cells (reticulocytes) by two different (semi-) automated flow-cytometry technologies

Flow cytometrical methods have been introduced recently as an alternative to the enumeration of reticulocytes by microscopy. Two of these methods have gained widespread use in haematological practice; the multiparametric flow cytometer using thiazole orange staining (Retic-Count, FACScan) and the single-application reticulocyte counter using auramine-O staining (R-series, Sysmex). Several studies have emphasized the excellent correlations between microscopy and these techniques. The purpose of our study has been to examine the specificity of these automated devices with regard to cells classified as 'reticulocytes' and the effect that this may have on measures of reticulocyte maturity. Our results indicate that the specificity of reticulocyte measurements by both the Sysmex R-1000/-3000 and the Retic-Count system is relatively low. This is due to the presence of leucocytes amongst cells classified as reticulocytes. These leucocytes display intense staining with either dye, leading to an erroneous estimation of RMI (thiazole orange) and high fluorescence count (R-1000/-3000). This error is directly correlated with the leucocyte count. The basis for reticulocyte identification should be improved before automated estimation of reticulocyte maturation can be used in clinical practice.

Impact of preleukapheresis cell counts on collection results and correlation of progenitor-cell dose with engraftment after high-dose chemotherapy in patients with germ cell cancer

PURPOSE

To identify predictive factors for a good leukapheresis yield and to determine peripheral-blood progenitor cell (PBPC) dose requirements for rapid hematopoietic engraftment.

PATIENTS AND METHODS

Seventy-one patients with germ cell cancer (GCC) underwent PBPC harvest for autologous transplantation following high-dose therapy. Aphereses were performed after chemotherapy during granulocyte colony-stimulating factor (G-CSF) administration.

RESULTS

A median of two aphereses (range, two to five) resulted in 4.6 x 10(8) mononuclear cells (MNC)/kg, 15.7 x 10(4) colony-stimulating units granulocyte-macrophage (CFU-GM)/kg, and 6.0 x 10(6) CD34+ cells/kg. Peripheral blood MNC count correlated significantly with number of harvested CD34+ cells per kilogram (r = .49; P < .0001) and with CFU-GM count per kilogram (r = .35; P < .002). Circulating CD34+ cells from peripheral blood gave the best correlations to collected CD34+ cells per kilogram (r = .92; P < .0001), as well as to harvested CFU-GM per kilogram (r = .48; P < .0001). A preleukapheresis number of CD34+ cells greater than 4 x 10(4)/mL was highly predictive for a PBPC collection yield that contained more than 2.5 x 10(6) CD34+ cells/kg harvested by a single leukapheresis. After autologous transplantation, 41 patients were assessable for hematopoietic engraftment. They engrafted in a median time of 9 days (range, 7 to 18) to a WBC count greater than 1.0 x 10(9)/L, 10 days (range, 7 to 18) to an absolute neutrophil count (ANC) greater than 0.5 x 10(9)/L, and 11 days (range, 7 to 62) to a platelet (PLT) count greater than 20 x 10(9)/L. Good correlations were seen between reinfused CD34+ cell count and recovery of WBC count, ANC, and PLT count, with r values of .65 (P < .001), .65 (P < .001), and .45 (P < .03), respectively. Patients reinfused with a PBPC dose greater than 2.5 x 10(6) CD34+ cells/kg recovered hematopoiesis in a significantly shorter time than patients who received less than 2.5 x 10(6) CD34+ cells/kg.

CONCLUSION

Rapid hematopoietic engraftment can be achieved by a PBPC dose of greater than 2.5 x 10(6) CD34+ cells/kg. When circulating preleukapheresis CD34+ cell counts are greater than 4 x 10(4)/mL, a PBPC autograft that contains more than 2.5 x 10(6) CD34+ cells/kg can be collected by a single leukapheresis.

Autografting with blood progenitor cells: predictive value of preapheresis blood cell counts on progenitor cell harvest and correlation of the reinfused cell dose with hematopoietic reconstitution

One hundred and nine patients suffering from various malignancies underwent 285 apheresis procedures for PBPC collection. A median of two leukaphereses (range: 2-5) resulted in median numbers of 4.6 x 10(8) MNC/kg, 14.1 x 10(4) CFU-GM/kg, and 6.0 x 10(6) CD34+ cells/kg. Preleukapheresis peripheral blood CD34+ cells correlated significantly with collected CD34+ cells/kg (r = 0.94; p < 0.0001) and with CFU-GM/kg (r = 0.52; p < 0.0001). A value > 4 x 10(4) CD34+ cells/ml was highly predictive for a collection yield > 2.5 x 10(6) CD34+ cells/kg harvested by a single leukapheresis. Sixty patients were evaluated for hematologic reconstitution and engrafted in a median time of 10 days for WBC > 1.0 x 10(9)/l (range: 7-21 days), 10 days for ANC > 0.5 x 10(9)/l (7-20) and 11 days for PLT > 20 x 10(9)/l (7-62). Reinfused CD34+ cells/kg correlated significantly with hematologic engraftment (r = 0.44-0.52 and p < 0.006-0.001) as well as CFU-GM/kg (r = 0.36-0.44 and p < 0.007-0.001). A progenitor cell dose > 2.5 x 10(6) CD34+ cells/kg or > 8.0 x 10(4) CFU-GM/kg led to a significantly faster recovery for WBC, ANC, and PLT when compared with patients receiving < 2.5 x 10(6) CD34+ cells/kg or < 8.0 x 10(4) CFU-GM/kg. We conclude that rapid hematopoietic engraftment after high-dose therapy and PBPC reinfusion correlates well with a progenitor cell dose > 2.5 x 10(6) CD34+ cells/kg or > 8.0 x 10(4) CFU-GM/kg, and that above a preleukapheresis threshold of 4 x 10(4) CD34+ cells/ml a PBPC autograft containing > 2.5 x 10(6) CD34+ cells/kg can be collected by a single leukapheresis. We suggest that patients recovering from myelosuppression should be monitored for CD34+ cells in serial blood samples to determine the course of circulating hematopoietic progenitor cells. This issue will help to define the optimal time point to start apheresis and to predict a PBPC autograft harvested by a single leukapheresis, which will lead to rapid and stable hematopoietic reconstitution following transplantation.

Propagation of large numbers of cells of a human mixed-lineage T-lymphoid/myeloid

Previously, a subset of T cells co-expressing the myeloid antigen CD33 has been described in patients with acute myelogenous leukaemia. However, normal lymphocytes have been viewed as not expressing the CD33 antigen. We have developed culture conditions which allow for the rapid expansion of CD3+CD33+ cells from patients with myeloid leukaemia as well as normal individuals. The protocol for cellular expansion includes the addition of interferon-gamma on day 0, interleukin-1, interleukin-2 and a monoclonal antibody against CD3 on day 1 to peripheral blood lymphocytes. Using this protocol, total cell number increased more than 600-fold within 16 d of culture. Cells could be kept in culture for more than 6 months. Cells of the CD3+CD33+ phenotype increased to 15.2 +/- 4.6% using this protocol after 16 d in culture. These cells have been characterized by flow cytometry and have been found to express the alpha, beta T-cell receptor, co-express the CD2, CD5, CD7 and HLA-DR antigens and did not express CD14 or CD15 antigens. Cells of the CD3+CD33+ phenotype were unable to lyse tumour cells as determined in a 51Cr release assay. In patients with chronic myeloid leukaemia. CD3+CD33+ cells seem to be negative for expression of bcr/abl transcript in contrast to CD33- cells. Our data suggest that CD3+CD33+ cells do exist in peripheral blood from normal individuals.

Hematopoietic rescue after high-dose chemotherapy using autologous peripheral-blood progenitor cells or bone marrow: a randomized comparison

PURPOSE

To compare autologous bone marrow (BM) with peripheral-blood progenitor cells (PBPC) as hematopoietic rescue after high-dose chemotherapy (HDCT).

PATIENTS AND METHODS

From January 1991 until April 1993, 47 consecutive patients with relapsed or refractory germ cell tumors were randomized to either BM harvest or collection of PBPC mobilized by chemotherapy plus granulocyte colony-stimulating factor (G-CSF). After additional conventional-dose salvage treatment, all patients received HDCT with carboplatin 1,500 mg/m2, etoposide 2,400 mg/m2, and ifosfamide 10 g/m2 with either BM or PBPC rescue.

RESULTS

Forty-six patients were assessable for hematologic reconstitution, and one patient died on day +4 before engraftment. Rescue using PBPC resulted in a significantly shorter recovery time to neutrophil counts more than 500/microL (10.0 v 11.0 days, P < .01), neutrophil counts more than 1,000/microL (10.0 v 12.0 days, P = .001), and platelet counts more than 20,000/microL (10.0 v 17.0 days, P < .01), as well as in fewer days to transfusion independence from RBCs (8.0 v 12.0, P < .05) and platelets (9.0 v 12.0, P < .01) and fewer days of intravenous (IV) antibiotics (9.0 v 11.0, P < .05). However, no statistical differences in transfusion requirements or in other clinical outcome variables were observed. Overall survival and event-free survival also were not different in the two study arms.

CONCLUSION

We conclude that the use of PBPC mobilized by chemotherapy plus G-CSF results in sustained trilineage reconstitution after HDCT, which occurs more rapidly as compared with BM. The earlier hematologic reconstitution in patients with PBPC rescue significantly reduces the time to transfusion independence.