Bone Marrow Transplantation, (1997) 20, 827–834  1997 Stockton Press All rights reserved 0268–3369/97 $12.00

CD34؉-selected autologous peripheral blood stem cell transplantation (PBSCT) in patients with poor-risk hematological malignancies and solid tumors. A single-centre experience

D Nachbaur1, F-M Fink2, W Nussbaumer3,AGa¨chter1, G Kropshofer1, C Ludescher4 and D Niederwieser1

1Division of Clinical Immunobiology and Bone Marrow Transplantation, Department of Internal Medicine, 2Department of Pediatrics, and 3Department of Transfusion Medicine, 4Department of Hematology-Oncology, University Hospital, Innsbruck, Austria

Summary: Peripheral blood stem cells (PBSC) are increasingly being used for treatment of hematological malignancies and solid Between July 1994 and December 1996, PBSC were tumors.1–8 Incomplete tumor eradication or malignant cells mobilized in 28 patients with poor-risk hematological contaminating the autograft contribute to treatment failure -malignancies and solid tumors. CD34؉ cells were posi- in up to 40–60% of patients after ABMT/PBSCT.9–13 Posi tively immunoselected using the Ceprate CS System. By tive selection of CD34+ cells as one approach to ameliorate December 1996, 22 patients had been reinfused with a these shortcomings results in a 1–4 log10 reduction of tumor -median of 3.325 (0.078–9.5) ؋ 106/kg CD34؉ cells. In cells in apheresis products, but whether this cost- and time three patients unselected back-up PBSC had to be consuming technique improves survival is currently under transfused along with selected CD34؉ cells because of a investigation.14–16 CD34؉ cell number Ͻ0.5 ؋ 106/kg. G-CSF (10 ␮g/kg) We report here on our experience in 28 patients suffering was started on day ؉1 and all patients engrafted within from poor-risk malignancies, for whom CD34+-selected a median day number of 12 (range, 10–22) until leuko- PBSC were collected between July 1994 and December -cytes Ͼ1.0 ؋ 109/l and a median day number of 56 1996. By December 1996, 22 patients had received an auto range, 10–180) until platelets Ͼ20.0 ؋ 109/l (ie platelet graft with CD34+-selected PBSC. Stem cell collection data) transfusion independence). Time to leukocyte and plate- are presented and clinical outcome is compared to a histori- let recovery was significantly shorter in patients receiv- cal control group autografted with unselected BM Ϯ PB .ing Ͼ2.0 ؋ 106/kg purified CD34؉ cells as compared to stem cells patients reinfused with Ͻ2.0 ؋ 106/kg CD34؉ cells. The hematopoietic recovery time was similar to that of 18 historical control patients treated with unseparated Patients and methods ABMT ؎ PBSCT with the exception of a significantly faster leukocyte engraftment in patients receiving Patients, cytapheresis and cryopreservation Ͼ2.0 ؋ 106/kg CD34؉ cells and a significantly delayed platelet recovery time in patients receiving Between July 1994 and December 1996, stem cells were (Ͻ2.0 ؋ 106/kg purified CD34؉ cells. There was a trend mobilized in 28 patients, 1–59 years old (median, 39 years for a better overall survival and a lower probability of with poor-risk hematological malignancies or solid tumors progression/relapse as compared to the historical con- with high-dose CY (7 g/m2, n = 19) or other dose-intensive trols. We observed five episodes of serious opportunistic chemotherapy regimens (n = 6) and rhG-CSF (10 ␮g/kg). infections (three pulmonary fungal infections, two cases By December 1996, 22 patients had been autografted with of cryptosporidiosis) after the take. Four of these purified CD34+ cells. The characteristics of these 22 patients had been reinfused with Ͻ2.0 ؋ 106/kg CD34؉ patients are listed in Table 1. All patients were considered cells probably indicating a delayed immune reconsti- to have poor-risk disease according to one of the following ,tution after CD34؉-selected PBSCT. criteria: multiple lines of treatment prior to transplantation Keywords: CD34+ selection; peripheral blood stem cell refractory or progressive disease implying Ͻ50% tumor transplantation; hematological malignancies; solid tumors reduction from initial treatment or Ͼ25% increase in tumor size, intermediate- or high-grade lymphoma achieving only partial remission after standard chemotherapy, solid tumors with metastatic disease at initial diagnosis or local or dis- seminated relapses, stage IIb to IV or inflammatory breast cancer. PBMC were collected in the recovery phase, when CD34+ cells in the peripheral blood reached values Ͼ0.3% Correspondence: Dr D Nachbaur, Division of Clinical Immunobiology and Bone Marrow Transplantation, Department of Internal Medicine, Anich- using a CS 3000 Plus blood cell separator (Baxter strasse 35, A-6020 Innsbruck, Austria Healthcare, Fenwal Division, Deerfield, IL, USA). Three Received 11 April 1997; accepted 13 July 1997 leukaphereses were performed for each individual patient. CD34؉-selected autologous PBSCT D Nachbauer et al 828 = cyclophos- = complete remission; PR (months) DX to TX to TX etoposide; CY = = melphalan, prednisone; IFN/P = ifosfamide, epirubicin, etoposide; = ifosfamide, methotrexate, etoposide; SIOP- = carboplatin; VP-16 radiotherapy; MP = = 3a), IFN 62 CR Ͼ epirubicin, cyclophosphamide; IEV = autologous bone marrow transplantation; CR = vincristine, doxorubicin, dexamethasone; RT = ifosfamide, doxorubicin, dactinomycin, vincristine; CP = plasma cell leukemia; ABMT = cyclophosphamide, doxorubicin, vincristine, prednisone; IMVP-16 = high-dose cyclophosphamide, doxorubicin, vincristine; COPP/ABVD, cyclophosphamide, vincristine, procarbazine, = Cooperative Study on Soft Tissue Sarcomas, protocol 91; EC -selected PBSC + = mitoxantrone, prednimustine; VAIA )) IV B IV refractory refractory COP, CHOP, IMVP-16 COP, CHOP, IFN 63 CR 19 PD = + + Austrian Neuroblastoma Study, protocol 94; VAD = angioimmunoblastic T cell lymphoma; PCL = cyclophosphamide, vincristine, prednisone; CHOP = (years) stable disease; A-NB-94 = Characteristics of 22 patients autografted with CD34 International Society of Pediatric Oncology, protocol 93-01; HD-CAV non-Hodgkin’s lymphoma; AILD cyclophosphamide, methotrexate, 5-FU; NOSTE = = = interferon/prednisone; COP phamide. CMF 93-01 prednisone/doxorubicin, bleomycin, vinblastine, DTIC; CWS-91 = partial remission; SD Table 1 UPN Sex153157 Age161163164 M171 M175 M177 F 3181 Diagnosis M 55186 M 42 Neuroblastoma190 Multiple myeloma F 45192 Multiple myeloma 52 F196 M Multiple 51 myeloma Multiple197 M myeloma High-grade200 1 NHL F (Ki1 22201 M Wilm’s 3 tumour202 Stage Hodgkin’s F 42 disease at diagnosis203 M Rhabdomyosarcoma 33204 High-grade M III NHL 33 A IV (Ki1 210 Breast III F cancer A 15 Nonseminomatous213 M germ 36 III cell Stage tumor A at216 M III Neuroblastoma TX A 2 Hodgkin’s M disease metastatic 42 diseaseNHL F Neuroblastoma 4 Intermediate-grade 59 F 3rd NHL IV relapse IV M after 51 Rhabdomyosarcoma ABMT Multiple myeloma 50 Multiple myeloma 51 Breast 23 PD cancer CR PR Follicular IIb NHL, CR SD Hodgkin’s Ewing disease II sarcoma IV Previous IV B, B treatment bulky refractory disease PCL IVB IV IV III A Interval III A CR CR PD Response PEB, COPP/ABVD, VAD IIb VAD, PEI ABMT, RT (bilateral) RT A-NB-94 MP metastatic (6 ( disease cycles) CR MP, PR RT, VAD PR CR very CR good PR PD 54 CR CWS-91 (8 SIOP-93-01, refractory CR RT, cycles), HD-CAV COPP/ABVD, RT IEV CR 7 EF A-NB 94 COP, (8 64 CHOP cycles) A-NB 9 94 29 (6 17 COPP/ABVD, cycles) CR COP, NOSTE, MP, IFN VAD CWS-96 21 25 (9 7 cycles) PD MP, VAD, CR VAIA, RT, CP/VP-16/CY, PD IFN CMF RT PR 101 relapse SD relapse CR 7 6 108 15 7 CR 43 CR CR 12 42 CR CR PR CR SD 6 CR CD34؉-selected autologous PBSCT D Nachbauer et al 829 The first two apheresis products were pooled and CD34+ fraction were stained without Ficoll–Isopaque centrifug- cells were positively immunoselected with a CD34-specific, ation. biotinylated MoAb (12–8) according to the manufacturer’s instructions using the Ceprate SC Stem Cell Concentration Transplant procedure, supportive care and maintenance System (CellPro, Bothell, WA, USA). Cells were cryopre- therapy served in 10% DMSO by control-rated freezing and stored in liquid nitrogen. The third apheresis product was cryopre- Patients were treated under strict reverse isolation without served unselected as back-up in all patients. laminar air-flow. No prophylactic systemic antibiotics were The historical control group consisted of 18 consecutive administered. All patients underwent a non-absorbable oral patients with poor-risk malignancies according to the gut decontamination with vancomycin, gentamycin and above-mentioned criteria autografted at our institution nystatin. Pneumocystis carinii prophylaxis was performed between 1983 and 1996 (solid tumor, n = 11; lymphoma, with trimethoprim-sulfamethoxazole given in a 10-day n = 3; multiple myeloma, n = 4). The two groups were not course before transplantation and after the take. Cytomega- different with respect to diagnosis, median age at trans- lovirus (CMV) pneumonia prophylaxis consisted of plant, median time to transplantation, lines of previous infusions of CMV hyperimmunoglobulin (1 ml/kg) every treatment, and median post-transplant follow-up (P = NS). other week until day +100. Irradiated (25 Gy), leukocyte- There were significantly more females in the historical con- depleted red cell and platelet transfusions from single trol group (11/18 vs 8/22, P = 0.0027) and significantly donors were administered when hemoglobin levels were 7.0 more patients in the control group received a TBI-contain- g/dl or less and platelets were 20 ϫ 109/l or less. ing conditioning regimen (10/18 vs 4/22 in the CD34+ Pretransplant conditioning regimens consisted of CY group, P = 0.0001). Patients received unselected hematopo- (120 mg/kg) and fractionated TBI (12 Gy) in two myeloma ietic stem cells from autologous BM plus PB (n = 11) or patients; busulphan (16 mg/kg) and CY (120 mg/kg) in four PB alone (n = 7). Patients who received BM plus PB stem myeloma patients; carboplatin (1500 mg/m2), etoposide (60 cells were reinfused with a median of 0.25 (range, 0.1– mg/kg) and melphalan (160 mg/m2) in seven patients with 0.37) ϫ 108/kg BM plus 0.95 (range, 0.18–4.34) ϫ 108/kg solid tumors; CY (100 mg/kg), BCNU (15 mg/kg) and eto- PB mononuclear cells. Patients who were autografted with poside (60 mg/kg) in four lymphoma patients; CY (100 PBSC alone received a median of 4.43 (range, 2.33– mg/kg), etoposide (30 mg/kg) and fractionated TBI (12 Gy) 7.66) ϫ 108/kg PB mononuclear cells. Twelve patients in two lymphoma patients; CY (120 mg/kg), carboplatin received G-CSF (10 ␮g/kg/day) beginning on day 1 post- (2000 mg/m2) and etoposide (1500 mg/m2) in one patient transplant to accelerate hematopoietic reconstitution (vs all with nonseminomatous germ cell tumor and carboplatin patients in the CD34+ group, P = 0.0002). (800 mg/m2), thiotepa (500 mg/m2), melphalan (90 mg/m2) and novantrone (30 mg/m2) in two breast cancer patients. Forty-eight hours after completing pretransplant condition- + CD34 quantitation ing, purified CD34 cells were rapidly thawed and given intravenously. All patients received G-CSF (10 ␮g/kg) Cell counts were performed using a fluorescence-activated starting on day +1 given as a 2-h infusion until leukocyte cell sorter (FACS Vantage; Becton Dickinson, Palo Alto, engraftment, defined as the first of 2 consecutive days with CA, USA). CD34+ cells were determined daily after stem leukocyte counts Ͼ1.0 ϫ 109/l (take). In three patients cell mobilization with high-dose chemotherapy in hep- additional unselected back-up PBMC were transfused arinized venous blood samples when leukocytes reached because of the low number of CD34+ cells values Ͼ1.0 ϫ 109/l, and in the leukapheresis products, the (Ͻ0.5 ϫ 106/kg). flow-through fraction from the CEPRATE column, and in Patients with multiple myeloma received recombinant the adsorbed fraction using a phycoerythrin (PE)-conju- interferon-␣-2b (Intron A; Schering International, Vienna, gated anti-human CD34 MoAb (HPCA-2; Becton Dickin- Austria) subcutaneously at a dosage of 2 mega units three son, Palo Alto, CA, USA). Peripheral blood mononuclear times weekly starting after stable engraftment, defined as cells (PBMC) from heparinized venous blood samples were leukocyte counts Ͼ3 ϫ 109/l and platelets Ͼ45 ϫ 109/l. obtained by Ficoll–Isopaque (Lymphoprep; Nycomed, Interferon treatment was maintained until disease pro- Oslo, Norway) gradient centrifugation (400 g for 30 s at gression. Patients with radiosensitive bulky or metastatic room temperature). After two washes cells were resus- disease at the time of autografting received involved field pended in phosphate-buffered saline (PBS) containing 1% irradiation within the first 3 months. bovine serum albumin and sodium azide and simul- taneously stained with anti-CD34 MoAb and fluorescein Response criteria (FITC)-conjugated anti-CD45 MoAb (HLe-1; Becton Dickinson) for 20 min at 4°C. Isotype-matched control anti- In patients with multiple myeloma, complete remission bodies were processed in parallel. After two washes, the (CR) was defined as less than 5% bone marrow plasma percentage of CD34+ cells was determined by gating on cells, no measurable paraprotein in serum and/or urine by forward (FSC) and side scatter (SSC) characteristics and immunoelectrophoresis or immunofixation, and no evi- CD45 staining. A minimum of 30 000 events were collected dence of progressive bony lesions. Partial response (PR) in list mode and processed with a CELLQuest software. was defined as у75% reduction of bone marrow plasma- MNC from the leukapheresis products, the flow-through cytosis and/or paraprotein levels without progressive bony fraction from the CEPRATE column and from the adsorbed lesions. Stable disease (SD) was defined as р75% CD34؉-selected autologous PBSCT D Nachbauer et al 830 reduction of bone marrow plasma cells and/or paraprotein TBI-containing conditioning regimen (UPN 177). No signs levels without new osteolytic lesions and progressive dis- of toxicity because of infusion of purified CD34+ cells were ease (PD) was defined as new osteolytic lesions and/or observed. All patients achieved leukocyte engraftment after у25% increase in marrow plasmacytosis and/or paraprotein a median of 12 (range 10–22) days and platelet transfusion levels.4 For patients with lymphoma or solid tumors CR independence after a median of 56 (range 10–180) days. was defined as the disappearance of all clinical and radio- As shown in Table 3, in patients receiving Ͻ2.0 ϫ 106/kg graphic evidence of disease. PR was defined as reduction CD34+ cells leukocyte engraftment was significantly of measurable disease by more than 50% without the delayed (median day number until take 15, range 11–22) appearance of new lesions, SD was defined as less than as compared to patients receiving Ͼ2.0 ϫ 106/kg CD34+ 50% tumor reduction and PD as the appearance of any new cells (median day number until take 11, range 10–13, tumor lesion and/or increase of measurable disease by more P = 0.0126). Time to platelet transfusion independence was than 25%. also significantly delayed in patients autografted with Ͻ2.0 ϫ 106/kg CD34+ cells (median days 118, range 16– 180) as compared to patients receiving Ͼ2.0 ϫ 106/kg Statistics CD34+ cells (median days 27, range 10–170, P = 0.0077, For the comparison of the distribution between the study Table 3). Both time to leukocyte engraftment and platelet and control group we used the ␹2 test for categorical vari- transfusion independence were significantly inversely cor- ables and the Wilcoxon test for continuous variables. Unad- related to the CD34+ cell number transfused (P = 0.008, justed Kaplan–Meier estimates were used to display the r =−0.6, and P = 0.0011, r = -0.7, respectively, Figure 1a incidences of overall survival, progression-free survival and and b). transplant-related mortality.17 Progression-free and overall The engraftment data of the CD34+ group of patients survival were defined as the time from the day of stem cell were compared with that of a historical group of patients transplantation (day 0) until disease progression or death. (n = 18) who received unmanipulated ABMT Ϯ PBSCT. Correlations between CD34+ cell number and time until Time to leukocyte or platelet recovery were not different take or platelet transfusion independence were computed for both groups with the exception of a significantly more using the Spearman rank correlation coefficient.18 For com- rapid leukocyte engraftment in patients autografted with parisons of time to engraftment between the independent Ͼ2.0 ϫ 106/kg purified CD34+ cells (median days until leu- groups we used the Wilcoxon rank-sum test.19 A P value kocytes Ͼ1.0 ϫ 109/l for the CD34+ group 11, range 10– р0.05 was considered significant. 13, vs 13, range 8–28, for the unselected control group, P = 0.0113) and a significantly delayed platelet recovery in patients receiving Ͻ2.0 ϫ 106/kg purified CD34+ cells as Results compared to unselected ABMT Ϯ PBSCT (median day number until platelets Ͼ20.0 ϫ 109/l, 118, range 16–180, CD34+ cell number and engraftment results for the CD34+ group vs 17, range 6–210, for the unselected group, P = 0.05). Twenty-eight patients had their circulating CD34+ cells pur- ified by the Ceprate SC concentrator. Apheresis products processing data are shown in Table 2. A median number Response, survival and causes of death of 6.0 (1.554–13.3) ϫ 108/kg MNC were processed with a CD34% ranging from 0.1–8% (median 1.5%). A median of All but five patients with measurable disease at the time of 4.28 (0.49–24.7) ϫ 106/kg adsorbed cells were obtained transplant showed an objective response, either partial or with a median purity of CD34+ cells after processing of complete, to high-dose myeloablative chemo/radiotherapy. 77.5% (range 16–95). Two patients had stable disease (UPN 197, 204) and three Twenty-two patients were reinfused with a median num- patients died of progressive disease 82, 267 and 416 days ber of 3.325 (0.078–9.5) ϫ 106/kg purified CD34+ cells. after the transplant, respectively (UPN 164, 192, 186). Two Three patients required additional reinfusion of unselected patients autografted in complete remission died of relapse back-up buffy coat cells because of the low CD34+ cell after 251 (UPN 175) and 384 days (UPN 181), respectively. number (Ͻ0.5 ϫ 106/kg). All three patients had multiple All other patients autografted in complete remission lines of treatment prior to transplantation with either remained progression-free throughout the observation per- alkylating agent-containing regimens for more than a year iod (Table 1). The median follow-up for the CD34+-selec- (UPN 161 and 164) or had a relapse after ABMT with a ted group was 10 months (range 3–28) and 8 months (range

Table 2 Apheresis products processing data in 28 patients with poor-risk hematological malignancies and solid tumors

Fraction MNC/kg CD34+ purity (%) CD34+/kg

Pre-selection 6.0 (1.554–13.3) × 108 1.5 (0.1–8) — Adsorbed 4.28 (0.49–24.7) × 106 77.5 (16–95) 3.1 (0.078–22.7) × 106 Unadsorbed — 0.2 (0–3) —

Results are given as median values (range). CD34؉-selected autologous PBSCT D Nachbauer et al 831 Table 3 CD34+ cell number and hematopoietic regeneration in 22 poor-risk patients autografted with purified CD34+ cells

6 6 + Ͼ2.0 × 10 /kg CD34+ cells Ͻ2.0 × 10 /kg CD34 cells P value reinfused (n = 13) reinfused (n = 9)

Days until leukocytes Ͼ1.0 × 109/l 11 (10–13) 15 (11–22) 0.0126 Days until platelet transfusion independence 27 (10–170) 118 (17–180) 0.0077 (Ͼ20.0 × 109/l)

Results are given as median values (range).

1–127) for the unselected historical control group a (difference statistically not significant). Overall survival for + 24 the CD34 -selected group was 57 and 16% for the historical controls (Figure 2a). The relapse/progression rate after r = –0.6 + 22 CD34 -selected PBSCT was 25% as compared to 75% of P = 0.008 the historical controls treated with unselected ABMT Ϯ PBSCT (Figure 2c). 20 Only one patient of the CD34+-selected cohort died from treatment-related complications (Figure 2b). This patient 18 (UPN 171) experienced intestinal cryptosporidiosis 52 days after reinfusion of 1.26 ϫ 106/kg CD34+ cells. Treatment 16 with paromomycin was unsuccessful. He developed pul- monary cryptosporidiosis and died of respiratory failure on + 14 day 118 while being in complete remission from lym- phoma. Four other patients developed serious opportunistic infections after the take (three pulmonary fungal infections, 12 Days until leukocytes >1.0 g/l UPN 190, day +60; UPN 201, day +21; UPN 203, day +32, and one Cryptosporidium infection, UPN 177, day +71) but 10 none of these patients died due to infection. Three of these four patients had also been reinfused with Ͻ2.0 ϫ 106/kg + 8 CD34 cells. 01 2 34 56 7 8 910 CD34+ cell number (×106/kg) Discussion b 200 Autologous transplantation with mobilized PBSC has been 180 r = –0.7 shown to be superior to BM stem cells with respect to P = 0.0011 engraftment kinetics but without any significant differences 160 in relapse incidence, transplant-related mortality or sur- vival.1,3,8,20 Interestingly, in one recent study an even higher 140 relapse incidence after unpurged PBSCT for lymphoma was reported as compared to antibody-purged ABMT.1 120 Concomitant tumor cell mobilization by the stem cell mobi- lization procedure has been noticed and tumor cells con- 100 taminating autografts have been identified by gene-marking 9–13 80 studies as a cause of relapse after ABMT/PBSCT. Posi- tive selection of CD34+ cells is one possibility to reduce 60 the quantity of contaminating tumor cells. However, the questions to be addressed with this technique include the Days until platelets >20.0 g/l 40 following: is it possible to mobilize a sufficient number of CD34+ cells even in extensively pretreated, poor-risk 20 patients, do CD34+-selected stem cells mediate a faster

0 hematopoietic regeneration than unseparated stem cells, 012 34 56 7 8910and is there any advantage with respect to overall survival, CD34+ cell number (×106/kg) transplant-related mortality or relapse incidence after CD34+-selected PBSCT. + Figure 1 Correlation between the number of reinfused CD34 cells/kg Using the Ceprate SC Stem Cell Concentration System and time to leukocyte engraftment (a) or platelet transfusion independence + (b). The lines represent median values of reinfused CD34+ cell number we obtained a sufficient number of CD34 stem cells even and of days until leukocyte recovery or platelet transfusion independence. in intensively pretreated poor-risk patients. All but one of + Correlations were computed using the Spearman correlation coefficient. nine patients in which Ͻ2.0 ϫ 106/kg CD34 cells were CD34؉-selected autologous PBSCT D Nachbauer et al 832 a obtained had received у3 lines of previous treatment 100 including either alkylating agents or irradiation or had bone marrow involvement of their underlying disease, factors that have all been identified as significant determinants of 80 progenitor cell yield.2,20,21 Additionally, three of these nine patients also experienced prolonged treatment with IFN-␣ prior to stem cell collection. Suppression of myeloid pro- + 60 57%, CD34 selected (n = 22) genitors by interferon treatment has been demonstrated and must be taken into account as an additional factor adversely affecting stem cell harvests.22 40 The rates of neutrophil and platelet recovery of the entire CD34+-selected group were equivalent when compared 16%, unselected (n = 18) with those of the historical control group. These results 20

Overall survival (% probability) confirm the observations made in two other small series using positively selected CD34+ cells for hematopoietic reconstitution after high-dose chemotherapy.14,23 Our obser- 0 vation of a significantly delayed platelet recovery in 0 1 2 3 4 5 6 7 8 9 10 11 patients autografted with Ͻ2.0 ϫ 106/kg purified CD34+ Years cells as compared to those receiving either a higher cell + b dose of selected CD34 or unselected BM Ϯ PB stem cells 24 40 is in line with the recently published data by Weaver et al about platelet reconstitution kinetics in patients autografted with unselected PBSC containing a low CD34+ cell dose and suggests that G-CSF alone is not sufficient to induce 30 a parallel increase of both granulocytes and platelets in this group of patients.24 Whether the use of post-transplant G- 18%, unselected (n = 18) CSF can even delay platelet recovery in patients receiving + 20 a low CD34 cell dose as it was demonstrated in a large retrospective analysis by the Seattle group remains to be shown in a prospective randomized trial.2 The third question of whether there is any survival 10 advantage for patients autografted with purified CD34+ 5%, CD34+ selected (n = 22) cells due to the tumor cell purging effect will not be answ- ered until controlled prospective randomized trials have 0 been completed. However, there was a trend towards a bet- Transplant-related mortality (% probability) Transplant-related 00234567891011 ter overall survival due to a lower relapse/progression prob- ability after CD34+-selected PBSCT. Whether ex vivo + Years expansion of purified CD34 cells resulting in reduction of contaminating tumor cells or combined positive and nega- c tive selection procedures could further improve the results 100 is under investigation.25–27 Furthermore, the stem cell mobi- lization procedure itself using high-dose CY might be a good prognostic parameter and might have contributed to 75%, unselected (n = 18) + 80 the beneficial effects of CD34 selection on overall survival and relapse incidence in our poor-risk patients. Another important aspect of CD34+ selection deals with 60 the concomitant T cell depletion and its possible conse- quences on immune reconstitution kinetics after CD34+- selected PBSCT, although purified, allogeneic CD34+ cells 40 devoid of mature T and B lymphocytes have been shown to restore lymphopoiesis in lethally irradiated non-human primates.28 Whether lymphocyte reconstitution after 25%, CD34+ selected (n = 22) + 20 CD34 -selected PBSCT is delayed or even is a function of CD34+ cell number resulting in a higher rate of infectious Relapse/progression (% probability) complications, as it is observed after T cell depletion in the allogeneic setting, is unknown.29 Preliminary observations 0 at our institution reveal a profound delay of qualitative and 02345678910111 quantitative T lymphocyte reconstitution, especially of + + Years CD4 cells, within the first 100 days after CD34 -selected PBSCT probably rendering those patients who receive a Figure 2 Probability of overall survival (a), transplant-related mortality low CD34+ cell dose (Ͻ2.0 ϫ 106/kg CD34+ cells) more (b) and relapse/progression (c) after CD34+-selected PBSCT (n = 22) and unselected ABMT Ϯ PBSCT (n = 18). CD34؉-selected autologous PBSCT D Nachbauer et al 833 susceptible to opportunistic infections even after successful origin of relapse after autologous bone marrow transplan- and prompt leukocyte engraftment.30 tation. Lancet 1993; 341: 85–86. In conclusion, CD34+ selection is feasible even in exten- 11 Rill DR, Santana VM, Roberts WM et al. Direct demon- sively pretreated poor-risk patients. Hematopoietic regener- stration that autologous bone marrow transplantation for solid ation was comparable to that of patients receiving unmanip- tumors can return a multiplicity of tumorigenic cells. Blood 1994; 84: 380–383. ulated BM and/or PB stem cells but, despite prompt 12 Brugger W, Bross KJ, Glatt M et al. Mobilization of tumor neutrophil regeneration, platelet recovery was significantly cells and hematopoietic progenitor cells into peripheral blood 6 + delayed in patients receiving Ͻ2.0 ϫ 10 /kg CD34 cells. of patients with solid tumors. Blood 1994; 83: 636–640. The survival data are encouraging but a prolonged follow- 13 Ross AA, Cooper BW, Lazarus HM et al. Detection and up and future, randomized studies are necessary to assess viability of tumor cells in peripheral blood stem cell collec- the effects of CD34+ selection on survival, relapse tions from breast cancer patients using immunocytochemical incidence and immune reconstitution. and clonogenic assay techniques. Blood 1993; 82: 2605–2610. 14 Lemoli RM, Fortuna A, Motta MR et al. Concomitant mobil- ization of plasma cells and hematopoietic progenitors into per- ipheral blood of multiple myeloma patients: positive selection Acknowledgements and transplantation of enriched CD34+ cells to remove circul- ating tumor cells. Blood 1996; 87: 1625–1634. This work was financially supported by grants of the Austrian 15 Schiller G, Vescio R, Freytes C et al. Transplantation of + Research Fund ‘Zur Fo¨rderung der wissenschaftlichen For- CD34 peripheral blood progenitor cells after high-dose schung’, project No. 10525. chemotherapy for patients with advanced multiple myeloma. Blood 1995; 86: 390–397. 16 Shpall EJ, Jones RB, Bearman SI et al. Transplantation of enriched CD34-positive autologous marrow into breast cancer References patients following high-dose chemotherapy: influence of CD34-positive peripheral-blood progenitors and growth fac- 1 Brunvand MW, Bensinger WI, Soll E et al. High-dose frac- tors on engraftment. J Clin Oncol 1994; 12: 28–36. tionated total-body irradiation, etoposide and cyclophospham- 17 Kaplan E, Meier P. Non-parametric estimation for incomplete ide for treatment of malignant lymphoma: comparison of auto- observations. J Am Stat Assoc 1958; 53: 457–481. logous bone marrow and peripheral blood stem cells. Bone 18 Brown BW. Hollander M. Linear regression and correlation. Marrow Transplant 1996; 18: 131–141. In: Brown B, Hollander M (eds). Statistics. Wiley: New York, 2 Bensinger W, Appelbaum F, Rowley S et al. Factors that 1977, pp 261–307. influence collection and engraftment of autologous peripheral- 19 Sachs L. Angewandte Statistik. Springer Verlag: Berlin, 1974. blood stem cells. J Clin Oncol 1995; 13: 2547–2555. 20 Tricot G, Jagannath S, Vesole D et al. Peripheral blood stem 3 Langenmayer I, Weaver C, Buckner CD et al. Engraftment of cell transplants for multiple myeloma: identification for favor- patients with lymphoid malignancies transplanted with auto- able variables for rapid engraftment in 225 patients. Blood logous bone marrow, peripheral blood stem cells or both. Bone 1995; 85: 588–596. Marrow Transplant 1995; 15: 241–246. 21 Dreger P, Klo¨ss M, Petersen B et al. Autologous progenitor 4 Bensinger WI, Rowley SD, Demirer T et al. High-dose ther- cell transplantation: prior exposure to stem cell toxic drugs apy followed by autologous hematopoietic stem-cell infusion determines the yield and engraftment of peripheral blood pro- for patients with multiple myeloma. J Clin Oncol 1996; 14: genitor cell but not of bone marrow grafts. Blood 1995; 86: 1447–1456. 3970–3978. 5 Weaver CH, Petersen FB, Appelbaum FR et al. High-dose 22 De Fabritiis P, Sandrelli A, Meloni G et al. Prolonged sup- fractionated total-body irradiation, etoposide, and cyclophos- pression of myeloid progenitor cell numbers after stopping phamide followed by autologous stem-cell support in patients interferon treatment for CML may necessitate delay in har- with malignant lymphoma. J Clin Oncol 1994; 12: 2559– vesting marrow cells for autografting. Bone Marrow Trans- 2566. plant 1990; 6: 247–251. 6 Elias AD, Ayash L, Anderson KC et al. Mobilization of per- 23 Brugger W, Henschler R, Heimfeld S et al. Positively selected ipheral blood progenitor cells by chemotherapy and granulo- autologous blood CD34+ cells and unseparated peripheral cyte–macrophage colony-stimulating factor for hematologic blood progenitor cells mediate identical hematopoietic support after high-dose intensification for breast cancer. Blood engraftment after high-dose VP16, ifosfamide, carboplatin, 1993; 79: 3036–3044. and epirubicin. Blood 1994; 84: 1421–1426. 7 Fukuda M, Kojima S, Matsumoto K, Matsuyama T. Auto- 24 Weaver CH, Potz J, Redmond J et al. Engraftment and out- transplantation of peripheral blood stem cells mobilized by comes of patients receiving myeloablative therapy followed chemotherapy and recombinant human granulocyte colony-sti- by autologous peripheral blood stem cells with a low CD34+ mulating factor in childhood neuroblastoma. Br J Haematol cell content. Bone Marrow Transplant 1997; 19: 1103–1110. 1992; 80: 327–331. 25 Willems P, Croockewitt A, Raymakers R et al. CD34 selec- 8 To LB, Roberts MM, Haylock DN et al. Comparison of tions from myeloma peripheral blood cell autografts contain haematological recovery times and supportive care require- residual tumour cells due to impurity, not to CD34+ myeloma ments of autologous peripheral blood stem cell transplants, cells. Br J Haematol 1996; 93: 613–622. autologous bone marrow transplants and allogeneic bone mar- 26 Widmer L, Pichert G, Jost LM, Stahel RA. Fate of contaminat- row transplants. Bone Marrow Transplant 1991; 9: 277–284. ing t(14;18)+ lymphoma cells during ex vivo expansion of 9 Deisseroth AB, Zu Z, Claxton D et al. Genetic marking shows CD34-selected hematopoietic progenitor cells. Blood 1996; that Ph+ cells present in autologous transplants of chronic 88: 3166–3175. myelogenous leukemia (CML) contribute to relapse after auto- 27 Di Nicola M, Bregni M, Siena S et al. Combined negative logous bone marrow in CML. Blood 1994; 83: 3068–3076. and positive selection of mobilized CD34+ blood cells. Br J 10 Brenner MK. Rill DR, Moen RC et al. Gene-marking to trace Haematol 1996; 94: 716–721. CD34؉-selected autologous PBSCT D Nachbauer et al 834 28 Andrews RG, Bryant EM, Bartelmez SH et al. CD34+ marrow 30 Nachbaur D, Kropshofer G, Feichtinger H et al. Cryptospori- cells, devoid of T and B lymphocytes, reconstitute stable lym- diosis after CD34-selected autologous peripheral blood stem phopoiesis and myelopoiesis in lethally irradiated allogeneic cell transplantation (PBSCT). Treatment with paromomycin, baboons. Blood 1992; 80: 1693–1701. azithromycin and recombinant human interleukin-2. Bone 29 Pirsch JD, Maki DG. Infectious complication in adults with Marrow Transplant 1997; 19: 1261–1263. bone marrow transplantation and T cell depletion of donor marrow. Increased susceptibility to fungal infections. Ann Intern Med 1986; 104: 619–631.