Transplantation (2001) 27, 683–687  2001 Nature Publishing Group All rights reserved 0268–3369/01 $15.00 www.nature.com/bmt CD34+ cell selection Combined positive and negative cell selection from allogeneic peripheral progenitor cells (PBPC) by use of immunomagnetic methods

GA Martı´n-Henao1, M Pico´n1, B Amill1, S Querol1, C Ferra`2, A Gran˜ena2 and J Garcı´a1

1Department of Cryobiology and Cell Therapy, Cancer Research Institute; and 2Hospital Dura´n i Reynals, Barcelona, Spain

Summary: The development of HLA-mismatched or haploidentical donor transplantation requires a high degree of T Twenty-four mobilized peripheral blood products from cell depletion (TCD) while preserving an adequate number healthy donors for allogeneic transplantation were posi- of CD34+ cells in the graft in order effectively to prevent tively selected for CD34+ cells and depleted of CD4+ and acute graft-versus-host disease (aGVHD) and ensure CD8+ cells (+/− selection) by combining clinical grade engraftment.1–3 immunomagnetic methods. A sequential, ‘two-step’ Several reasons make CD34 positive cell selection strategy combining positive selection of CD34+ cells by methods (positive selection, + selection) particularly attract- use of the Isolex 300i (versions 1 and 2) device and T ive to deplete T cells especially in peripheral blood progeni- cell depletion (TCD) using the MaxSep device and a tor cell (PBPC) grafts: they can be applied to the large simultaneous, ‘one-step’ method of CD34+ cell selection numbers of cells present in PB, they are able to preserve and TCD using the Isolex 300i (software versions 1 and adequate numbers of progenitors after selection, thus main- 2) have been investigated. Using these magnetic bead taining the potential benefits of a PBPC transplant in this separation systems, two groups of sequential +/− selec- setting and they passively deplete the elevated numbers of tion (Isolex 300i version 1/MaxSep and Isolex 300i ver- T cells present in the PB grafts. sion 2/MaxSep) and two groups of simultaneous +/− Several methods are available for positive selection.4–7 selection (Isolex 300i versions 1 and 2) were analysed. The immunomagnetic technique, which is now exploited In the sequential +/− selection, logarithms of TCD by the Isolex 300i cell selection technology (Nexell Thera- (CD3+ cell depletion) obtained by the positive selection peutics, Irvine, CA, USA), involves specific binding of tar- step had median values of 3.7 with the version 1 (n = 5) get cells by a mouse anti-CD34 MoAb, 9C5. The target and 4.5 with version 2 software of the Isolex 300i (n = 5) cells–9C5 complexes are then captured by sheep anti- (P = 0.07). Version 2 also gave a higher CD34+ cell mouse IgG-coated paramagnetic beads (IB) (SAM-M450; purity and yield than did version 1 (92% vs 77%, Dynabeads, Dynal, Oslo, Norway) and in a final step target P Ͻ 0.05 and 55% vs 34%, P = 0.3, respectively). cells are released from the 9C5-IB complexes by a peptide Additional TCD obtained in the second step with the that binds competitively to the 9C5 MoAb.8 This method MaxSep device for the two groups had a median value achieves between 3 and 4 logs of TCD,9 thus resulting in of 0.9 log and 7% CD34+ cell losses. In the simultaneous doses in the allograft lower than those suggested to +/− selection, the Isolex 300i version 2 (n = 10) gave a trigger aGVHD in the T cell-depleted BM of the HLA- median TCD of 5.1 log and version 1 (n = 4) of 4 log matched setting.10–13 However, in recipients of HLA- (P Ͻ 0.005). Higher CD34+ cell purity and yield were mismatched or haploidentical donor grafts, a higher TCD also obtained with version 2 than with version 1 (97% of the graft may be required to avoid aGVHD.2,3 Methods and 76%, P Ͻ 0.005 and 57% and 39%, P = 0.07, that allow additional T cell removal from the purified respectively). These data indicate that simultaneous, CD34+ cell product combining both CD34 positive and T ‘one-step’ +/− selection in the Isolex 300i version 2 achi- negative cell selections (positive/negative selection, +/− eves a high TCD with a high CD34+ cell purity and an selection) are of clinical relevance. In this sense, CD34+ acceptable CD34+ cell yield. Bone Marrow Transplan- cell-enriched components obtained with the Isolex 300i tation (2001) 27, 683–687. device leave cell surface antigens intact and free of surface- Keywords: CD34 positive and CD4 CD8 negative selec- bound reagent and allow additional rounds of negative tion; peripheral blood progenitor cells; Isolex 300i device cell selection. Here, we analyse the effectiveness of combined CD34 positive and T cell negative immunomagnetic selection strategies (+/− selection) for PBPC in the allogeneic setting; Correspondence: Dr GA Martı´n-Henao, Department of Cryobiology and + − Cell Therapy, Cancer Research Institute, Hospital Duran i Reynals, Av. a sequential, ‘two-step’ / selection (Isolex 300i for CD34 Castelldefels Km 2,7, L’Hospitalet de Llobregat, 08907 Barcelona, Spain positive selection and the Max-Sep device for T cell nega- Received 24 July 2000; accepted 18 October 2000 tive selection) has been investigated and a new strategy of Positive–negative cell selection in allogeneic PBPC GA Martı´n-Henao et al 684 simultaneous, ‘one-step’ +/− selection by the use of the Iso- 4°C, on end-over-end rotation at 4 r.p.m. Rosetted target lex 300i device (for both, CD34 positive and T cell negative cells were then captured by exposure to the magnet of the selection) has been analyzed. Potential improvements achi- MaxSep device (Baxter Healthcare) following the eved with the newly developed software for the Isolex 300i recommendations of the manufacturer. (version 2) have been compared with the previous system (version 1). Simultaneous ‘one-step’ +/− selection with Isolex 300i Software version 1: A combined approach of simultaneous Materials and methods +/− cell selection can be easily adapted to the Isolex 300i version 1 system following the manufacturer’s indications. Donors In some procedures (n = 4) a +/− selection was performed as follows: during the procedure of positive selection step Twenty-four fresh PBPC products from 14 healthy donors and following the transfer of the peptide release agent to collected after mobilization of hematopoietic stem cells the primary chamber, CD4-IB and CD8-IB complexes were with glycosylated rhG-CSF were processed. In 11 cases the added to it and gently mixed. These complexes would bind donor was an HLA-mismatched sibling and in three cases to the residual contaminating T cells and retain them in the an HLA-identical sibling. magnetic field (negative selection step); CD34+ cells will Leukaphereses (LK) obtained with a COBE Spectra then be released from the 9C5 MoAb–IB complexes into (COBE, Lakewood, CO, USA) or a CS3000 (Fenwal, the fluid phase. Concentration of IB–MoAb complexes in Baxter Healthcare, Deerfield, IL, USA) devices were the primary chamber was approximately 10 × 106/ml. The × 9 started when the WBC count in PB exceeded 1.0 10 /l rest of the process was allowed to continue with the + × 3 and the CD34 cell count 20 10 cells/ml. All donors had automated machine. given written informed consent and the study had been previously approved by the local ethics and research Software version 2: The up-grade software of the Isolex committees. 300i machine (version 2.0) has been adapted to perform all steps of combined +/− cell selection automatically. The Combined +/− cell selection procedures CD4 and CD8 MoAb (0.5 mg of each) and the IB (4 × 109) reagents were introduced into the primary chamber of the + + Working buffer (PBA) consisted of Ca2 /Mg2 -free PBS machine at different times when indicated by the device. (Baxter) with 1% (w/v) HSA and 12% (v/v) sodium citrate. The additional negative selection step was done as follows: LKs were incubated with 0.5% (w/v) human immunoglob- during the last wash of the CD34+ cells-9C5 MoAb–IB ulin (Grifols, Barcelona, Spain) for 15 min at room tem- complexes from the positive selection step, an additional perature (RT) without shaking. During this time, cellular incubation with CD4 and CD8 MoAb would bind to the aggregates were removed by filtering the cells through a residual contaminating T cells (negative selection step). 170 ␮m blood filter (Baxter) and collected into a 1000 ml After a mixing phase, the procedure continues until the Lifecell bag (Baxter) before transfer into the Isolex 300i transfer of the release peptide to the chamber takes place; device. The IB (Dynal) were washed twice with PBA at that time, additional IB are added. CD34+ cells are before addition to the system. released from the 9C5 MoAb–IB complexes into the fluid phase, the additional IB bind the target cells of the negative Sequential ‘two-step” +/− cell selection (Isolex 300i for selection step and retain them in the magnetic field. positive and Max-Sep for negative selection) Cell viability was evaluated by Trypan Blue dye exclusion in the LK products and in the final fractions. All Isolex 300i for positive selection: Separation procedures cell counts were performed using a Coulter JT3 counter were done by means of the full automatic Isolex 300i (Coulter Electronic, Hialeah, FL, USA). device. To date, two major modifications of the device software have been made (version 1 and 2). Five selection Cytometry analyses procedures were done with the version 1 and five with the version 2. The manufacturer’s recommendations were fol- Samples from LK products before separation, CD34+ cell- lowed with minor modifications as reported elsewhere.9 enriched, CD34+ cell-enriched T cell-depleted and CD34+ cell-depleted components were obtained for CD34+ cell Max-Sep for negative selection: A negative selection step enumeration as previously reported in detail.9 Briefly, cells was performed afterwards in an attempt to optimize the were incubated with human AB serum and anti-CD34-PE TCD to that achieved with the positive selection process. MoAb (8G12, HPCA-2; Becton Dickinson, San Jose, CA, The CD34+ cell-enriched component obtained with the Iso- USA) for 15 min at RT in the dark. After washing twice, lex 300i was transferred to a Lifecell bag (Baxter). Without the samples were incubated for 10 min at RT with further manipulation, the cell suspension was incubated ammonium chloride (Ortho Lysing Reagent; Ortho, Rari- with CD4 and CD8 MoAbs previously bound to IB. CD4- tan, NJ, USA) for erythrocyte lysis. Cells were then incu- IB and CD8-IB complexes were added to the cell suspen- bated with 7-aminoactinomycin-D (7-AAD; Molecular sion to give a final concentration of 10 × 106 of CD4-IB Probes, Eugene, OR, USA) for 5 min at RT. For enumer- and CD8-IB/ml, which had been found optimal in preclini- ation of CD34+ cells an amorphous gate identifying low cal assays (data not shown), and incubated for 30 min, at side scatter, bright CD34+ cells was drawn on a dot-plot of

Bone Marrow Transplantation Positive–negative cell selection in allogeneic PBPC GA Martı´n-Henao et al 685 side scatter and FL2 fluorescence, after excluding 7-AAD+ cell depletions in the first step of the procedure were 4.6 dead cells and including all NC. The absolute number of log (3.4–4.8) and 4.5 log (3.5–4.8), respectively, higher CD34+ cells was obtained by multiplying the percentage of than those achieved with the version 1 (P = 0.07). CD34+ viable cells by the cell count and the volume of the Additional CD4+ and CD8+ cell depletions performed with sample. The following panel of MoAb in single and double the MaxSep device in the CD34+ cell-enriched component combinations was used: CD3 (PE, SK7) from Becton gave final depletions of 5.1 log (4.7–5.9) and 5.1 log (4.1– Dickinson, CD4 (PE, SFCI12T4C11), CD8 (FITC, 5.3), respectively. The median increases of TCD (log) in SFCI21Thy2D3), CD19 (FITC, 89B) from Izasa-Coulter the negative selection step for CD4 and CD8 were 0.7 (0– (Izasa, Barcelona, Spain). Flow cytometry was performed 1.4) and 0.6 (0.2–0.8). CD34+ cell yield during the second on an Epics-XL-MCL cytometer (Coulter Electronic) step of negative selection was 99% (82–112). equipped with a 488 nm excitation Argon laser. Simultaneous ‘one-step’ +/− selection Clonogenic assays Isolex 300i version 1 (n = 4): NC load and percentage of Clonogenic assays were performed in PBPC before separ- starting CD34+ cells in the Isolex 300i version 1 were ation, CD34+ cell-enriched, CD34+ cell-enriched T cell- 4.7 × 1010 (2.9–5.5) and 1% (0.6–1.1), respectively. Logs depleted and CD34+ cell-depleted components. A commer- of CD4+ and CD8+ cell depletions were 4.2 (4–4.5) and 3.8 cially prepared methylcellulose-based medium (Methocult (3.8–4.1), and CD34+ cell purity and yield in the double GF 4434; Stem Cell Technologies, Vancouver, Canada) selected component were 76% and 39%, respectively. was used as previously described.14 Assays were run in trip- ° = licate. Cultures were incubated at 37 C and 5% CO2. Col- Isolex 300i version 2 (n 10): NC load and percentage onies were enumerated using an inverted microscope after of starting CD34+ cells in the Isolex 300i version 2 were 14 days of culture. 3.2 × 1010 (1.8–6.5) and 0.97% (0.7–1.7), respectively. Logs of CD4+ and CD8+ cell depletions were 5 (4.5–5.6) and 4.9 (4.2–5.5), respectively, higher than those obtained Statistical analyses with version 1 (P Ͻ 0.005). CD34+ cell purity and yield in Data are shown as medians and ranges unless otherwise the final, double-selected component were 97% and 57%, indicated. Continuous variables were compared using the respectively, also higher than with version 1 (P Ͻ 0.005 Mann–Whitney U test or Wilcoxon test. A P value and P = 0.07). lower than 0.05 was considered significant. Statistical analyses were performed using the SPSS 7.5 software Clonogenic assays (SPSS, Chicago, IL, USA). Recoveries of CFU-GM in the CD34+ cell-enriched compo- nents after + and +/− selection steps were 35% (24–55) and Results 22% (8–199), respectively (n = 8). Recovery of CFU-GM in the double-selected product was 18% (3–70) (n = 14) for + − + Processing simultaneous / selection procedures and in the CD34 cell-depleted component was 6% (1–58) (n = 22) for all All the selection procedures on a clinical scale were done procedures. Clonogenic efficiencies (for all CFU-GM, from a single LK. The median number of LK per donor BFU-E and CFU-Mix) per CD34+ cell in the initial, CD34+ was one (range 1–3). Results are shown in Tables 1 and 2. cell-enriched, CD34+ cell-enriched/T cell-depleted and CD34+ cell-depleted components were 27% (7–124), 10% Sequential ‘two-step’ +/− selection (6.5–20), 14% (6.5–27) and 27% (6–119), respectively. Cell viability as determined by Trypan blue dye exclusion Isolex 300i version 1 (n = 5): NC load and percentage of was 99% (93–100) in the initial material and 96% (88–99) initial CD34+ cells in the Isolex 300i version 1 were in the final double-selected product. 5 × 1010 (3.9–8) and 0.72% (0.4–1.2), respectively. CD4+ and CD8+ cell depletions in the positive selection (first step) were 3.7 log (3–3.8) and 3.7 log (3–3.8), respectively. Discussion Additional CD4+ and CD8+ cell depletions performed with the MaxSep device in the CD34+ cell-enriched component In the present study we analyzed the effectiveness of com- gave a final depletion of 4.8 log (4.5–5.8) and 4.3 log (3.9– bined CD34 positive and T cell negative selection on PBPC 5.6), respectively. The median increases of TCD (log) (= for allogeneic transplantation using different immuno- TCD double selection (log) − TCD first selection step (log)) magnetic strategies. in the negative selection step for CD4 and CD8 were 1.4 The results obtained with a simultaneous +/− cell selec- (1.1–2.0) and 0.9 (0.5–1.8). CD34+ cell yield during the tion with the Isolex 300i device version 2 with a median 5 second step of negative selection was 93% (86–106). log TCD, nearly 60% CD34+ cell yield and more than 90% of CD34+ cell purity, compare favorably with those Isolex 300i version 2 (n = 5): NC and percentage of CD34+ obtained with the previous software version of this cells loaded in the Isolex 300i version 2 were 4.7 × 1010 machine. Moreover, when the results obtained with the (3.5–7.8) and 0.9% (0.7–0.9), respectively. CD4+ and CD8+ positive selection step were analyzed, a higher CD34+ cell

Bone Marrow Transplantation Positive–negative cell selection in allogeneic PBPC GA Martı´n-Henao et al 686 Table 1 Combined CD34 positive and CD4/CD8 negative selection in PBPC allografts

Software version Two steps One step

1(n= 5) 2 (n = 5) 1 (n = 4) 2 (n = 10) After + After +/− After + After +/− After +/− After +/−

CD34+ cell purity (%) 77a 77.5 92a 96 76d 97d (64–91) (63–92) (88–98) (92–97) (62–92) (91–99) CD34+ cell yield (%) 34b 29 55.5b 46 39e 57e (16–61) (15–56) (23–64) (21–73) (27–58) (39–68)

c c d d TCD (log10) 3.7 4.5 4.5 4.6 4 5.1 (2.9–3.9) (4.1–5.0) (3.4–4.8) (4.4–5.3) (3.8–4.2) (4.4–5.6) ⌬ TCD (log10) NA 1.1 NA 0.3 NA NA (0.6–1.2) (0.0–1.1)

Two-step CD34+/CD4−CD8− selections were done with the Isolex 300i (version 1 or 2, positive selection step) followed by the Max-Sep (additional negative selection step) devices. One step CD34+/CD4−CD8− cell selection was performed with the Isolex 300i device (software version 1 or 2). TCD + (log10) = CD3 cell depletion. ⌬ TCD (log10) = TCD (double selection) (log10) − TCD (first selection step) (log10). aP Ͻ 0.05; bP = 0.3; cP = 0.07; dP Ͻ 0.005; eP = 0.07. NA = not applicable.

purity and yield and a higher rate of TCD, in the order of 4.5 log, were obtained with version 2 with respect to ver- sion 1. On the other hand, as the final CD34+ cell compo- Table 2 Percentage of T lymphoid cell subsets in the different cellular components of allogeneic PBPC after a double CD34 positive/CD4 and nent obtained with this technique is free of the residual CD8 negative selection reagents used during the procedure, an additional purging effect may be achieved by combining it with a second nega- Two steps One step tive immunomagnetic step if the T cell content of the graft is higher than is desirable. In this sense, our processing Version 1 Version 2 Version 1 Version 2 protocol of negative selection with the Max-Sep device = = = = (n 5) (n 5) (n 4) (n 10) may increase the TCD by about 0.5–1 log compared to that obtained with the positive selection procedure alone with Starting components + CD3+ (%) 40.8 30.5 31 34.3 minimal additional CD34 cell losses. (17.9–55) (24–46.6) (26.5–47.2) (25.7–46.6) Different strategies have been used to deplete T cells CD4+ (%) 23.6 14.6 21 19.2 from allo-PBPC grafts. Dreger et al15 used a negative cell (11.8–26.3) (13.8–15.5) (14.3–38.5) (14.7–31.6) + selection approach with Campath-1 MoAb and complement CD8 (%) 13.8 10.5 11 15.5 + (7–26.3) (9.8–11.3) (11.1–18.6) (9.1–22.8) and obtained about 2 log TCD and a 56% CD34 cell recov- CD19+ (%) 7.8 9.9 13.4 4.8 ery. Rambaldi et al16 described a two-step negative selec- (7.3–13) (8.5–14.4) (11–21.3) (3.7–9.4) tion by incubation with CD2 and CD7 magnetic microbeads and separation with magnetic field depletion columns after CD34-enriched components CD3+ (%) 2.4 0.18 NA NA debulking the graft of mature cells by an immune rosetting + (1.4–8.1) (0.1–3.4) technique. They achieved 3 log TCD with a CD34 cell CD4+ (%) 1.2 0.06 NA NA yield of 70%. Nevertheless, these approaches result in T (0.1–4.8) (0.06–2.3) + cell numbers in the graft that may still trigger GVHD, parti- CD8 (%) 0.7 0.07 NA NA 2,3 (0.5–3.1) (0.05–1) cularly in transplants across HLA barriers. A new approach to reduce the high numbers of T cells present in CD34-enriched/CD4-CD8-depleted components + + the allo-PBPC grafts is CD34 cell selection. Various CD3 (%) 0.37 0.1 0.6 0.04 methods have been developed for CD34+ cell purification (0.1–0.7) (0.03–0.3) (0.4–0.9) (0.01–0.2) + CD4+ (%) 0.1 0.06 0.25 0.03 in the clinical setting. The biotin–avidin-mediated CD34 (0.01–0.2) (0.01–0.1) (0.2–0.4) (0.00–0.08) cell selection method (Ceprate system) achieves about 3 log CD8+ (%) 0.2 0.03 0.35 0.02 TCD and 60–70% CD34+ cell purity and yield.7 In order to (0.01–0.7) (0.00–0.3) (0.2–0.4) (0.01–0.1) + increase the TCD with this technique, some authors have CD19 (%) 6 0.9 18.8 0.9 + (4.7–25.3) (0.3–6.2) (3.3–35.5) (0.09–4.2) applied a T cell-depletion step after selection of CD34 cells. For instance, Bertolini et al17 have tested a step of Results are shown as median and range. Sequential (two-step) negative selection with anti-CD3 MoAb and magnetic col- CD34+/CD4−CD8− selections were done with the Isolex 300i (version 1 loidal particles by using the StemSep device to deplete T or 2) (positive selection step) followed by negative selection with the Max- cells from the CD34+ cell-enriched component obtained Sep device. Simultaneous (one-step) selections were performed with the Isolex 300i (software version 1 or 2). with the Ceprate system in laboratory-scale samples. They + NA = not applicable. reported a 4.5 log TCD with about 50% of CD34 cell

Bone Marrow Transplantation Positive–negative cell selection in allogeneic PBPC GA Martı´n-Henao et al 687 recovery. Similar results were obtained by Martin-Hernan- 6 Cancelas JA, Querol S, Canals C et al. Peripheral blood dez et al18 with an anti-CD2 biotinylated MoAb and immu- CD34+ cell immunomagnetic selection in breast cancer noadsorption depletion columns after positive selection patients: effect on hematopoietic progenitor content and hema- using the same system. However, the Ceprate system is no tologic recovery after high-dose chemotherapy and autotrans- longer available for this purpose. Schumm et al19 obtained plantation. Transfusion 1998; 38: 1063–1070. + 7 Urbano-Ispizua A, Solano C, Brunet S et al. Allogeneic trans- a CD34 cell yield of 60–70% and a purity greater than + plantation of selected CD34+ cells from peripheral blood: 90% with a residual CD3 cell content in the enriched experience of 62 cases using immunoadsorption or immunom- component below 1% by using a magnetic field separation agnetic technique. Spanish Group of Allo-PBT. Bone Marrow device (Clinimacs system). Similar results were obtained Transplant 1998; 22: 519–525. by McNiece et al,20 who reported 4 to 5 log TCD with 8 Tseng-Law J, Szalay P, Guillermo R et al. Identification of a the same system. It should be noted that magnetic particles peptide directed against the anti-CD34 antibody, 9C5, by remain bound to the CD34+ cells in the enriched component phage display and its use in selection. obtained with this system, thus raising doubts about the Exp Hematol 1999; 27: 936–945. effectiveness of additional magnetic purging steps if the T 9 Martı´n-Henao GA, Pico´n M, Amill B et al. Isolation of + cell content of the purified component is higher than is CD34 progenitor cells from peripheral blood by use of an desirable. automated immunomagnetic selection system: factors affect- + ing the results. Transfusion 2000; 40: 35–43. The double-selected CD34 /T cell-depleted component 10 Atkinson K, Farrelly H, Cooley M et al. Human marrow T showed normal in vitro proliferation. Recovery of CFU- cell dose correlates with severity of subsequent acute graft- GM and clonogeneic efficiency after the second step of versus-host disease. Bone Marrow Transplant 1987; 2: 51–57. negative selection was similar to that obtained after positive 11 Kernan NA, Collins NH, Juliano L et al. Clonable T lympho- selection alone indicating minimal clonogeneic cell losses cytes in T cell-depleted bone marrow transplants correlate after the additional negative selection step. with development of graft-v-host disease. Blood 1986; 68: In conclusion, our data indicate that the simultaneous, 770–773. one-step +/− selection method with the Isolex 300i version 12 Lowenberg B, Wagemaker G, van Bekkum DW et al. Graft- 2 device is an effective technique able to obtain high versus-host disease following transplantation of ‘one log’ ver- degrees of TCD with a high CD34+ cell purity and an sus ‘two log’ T-lymphocyte-depleted bone marrow from acceptable yield. This would result in a final graft with very HLA-identical donors. Bone Marrow Transplant 1986; 1: low numbers of T cells and high numbers of CD34+ cells 133–140. 13 Verdonck LF, Dekker AW, de Gast GC et al. Allogeneic bone that might avoid severe aGVHD while preserving rapid marrow transplantation with a fixed low number of T cells in engraftment in the HLA-mismatched allogeneic setting. the marrow graft. Blood 1994; 83: 3090–3096. 14 Querol S, Capmany G, Azqueta C et al. Direct immunomag- netic method for CD34+ cell selection from cryopreserved Acknowledgements cord blood grafts for ex vivo expansion protocols. Transfusion 2000; 40: 625–631. 15 Dreger P, Viehmann K, Steinmann J et al. G-CSF-mobilized This work has been partially supported by grant SAF98-0050. peripheral blood progenitor cells for allogeneic transplan- Reagents used in the procedures of simultaneous positive– tation: comparison of T cell depletion strategies using different negative selection were kindly supplied by Baxter Healthcare. CD34+ selection systems or CAMPATH-1. Exp Hematol 1995; 23: 147–154. 16 Rambaldi A, Borleri G, Dotti G et al. Innovative two-step References negative selection of granulocyte colony-stimulating factor- mobilized circulating progenitor cells: adequacy for autolog- 1 Ringden O, Nilsson B. Death by graft-versus-host disease ous and allogeneic transplantation. Blood 1998; 91: 2189– associated with HLA mismatch, high recipient age, low mar- 2196. row cell dose, and splenectomy. Transplantation 1985; 40: 17 Bertolini F, Thomas T, Battaglia M et al. A new ‘two step’ 39–44. procedure for 4.5 log depletion of T and B cells in allogeneic 2 Aversa F, Tabilio A, Velardi A et al. Treatment of high-risk transplantation and of neoplastic cells in autologous transplan- acute leukemia with T-cell-depleted stem cells from related tation. Bone Marrow Transplant 1997; 19: 615–619. donors with one fully mismatched HLA haplotype. New Engl 18 Martı´n-Hernandez MP, Arrieta R, Martı´nez A et al. Haploid- J Med 1998; 339: 1186–1193. entical peripheral blood stem cell transplantation with a com- 3 Muller S, Schulz A, Reiss U et al. Definition of a critical T bination of CD34 selection and T cell depletion as graft- cell threshold for prevention of GVHD after HLA non- versus-host disease prophylaxis in a patient with severe com- identical PBPC transplantation in children. Bone Marrow bined immunodeficiency. Bone Marrow Transplant 1997; 20: Transplant 1999; 24: 575–581. 797–799. 4 Berenson RJ, Bensinger WI, Hill RS et al. Engraftment after 19 Schumm M, Lang P, Taylor G et al. Isolation of highly pur- infusion of CD34+ marrow cells in patients with breast cancer ified autologous and allogeneic peripheral CD34+ cells using or neuroblastoma. Blood 1991; 77: 1717–1722. the CliniMACS device. J Hematother 1999; 8: 209–218. 5 Martı´n-Henao GA, Ingle´s-Esteve J, Cancelas JA, Garcia J. 20 McNiece I, Briddell R, Stoney G et al. Large-scale isolation Isolation of CD34+ hematopoietic progenitor cells in chronic of CD34+ cells using the Amgen cell selection device results myeloid leukemia by magnetic activated cell sorting (MACS). in high levels of purity and recovery. J Hematother 1997; 6: Bone Marrow Transplant 1996; 18: 603–609. 5–11.

Bone Marrow Transplantation