Journal of Clinical Apheresis 7:lOl-109 (1992)
A Fully Automated Method for Mononuclear Bone Marrow Cell Concentration J.M. Rodriguez, M. Carmona, P. Noguerol, M. Ruiz, R. Parody, J.M. Perez-Hurtado, and I. Espigado
F. Vidal,
Department of Hematology, Virgen del Rocio University Hospital, Seville, Spain (J.M. R., M.C., P.N., M.R., R.P., J.M.P.-H., I.E.) and Baxter, Spain (F.V.) We describe our experience in processing 40 bone marrow aspirates harvested for autotransplantation from patients with several hematological diseases using the CS-3000 blood cell separator. The bone marrow of the first 30 patients was processed by a semiautomated method, and a fully automated procedure was used for the remaining 10 cases. Both procedures were developed in our laboratory and yielded a similar average mononuclear cell recovery of 87.78% and 86.988, respectively, and similar nucleated cell recovery (27.39% and 27.11%). The cloning efficiency of hematopoietic progenitor cells, measured as the total CFU-GM colony recovery in the in vitro cultures, did not differ between processed and recovered mononuclear cells. On the other hand, all the patients with transplants showed complete hematologic recovery, and the time to engraftment was similar to that described for other procedures. The automated procedure resulted in an average red cell removal of 97.81%, similar to the semiautomated procedure (94.198), though with a narrower range (96.31-98.6% vs. 80.34-98.34%). The time taken to process a similar amount of bone marrow cell suspension was very different for each method: 1 hour for the fully automated vs. 2’12 hours for the semiautomated method to process 1,000 ml. Furthermore, the semiautomated procedure required the addition of homologous or irradiated plasma in a laminar air flow chamber, while the automated method is performed in a closed sterile system. We conclude that our procedure using the CS-3000 processor is an efficient method for fully automated large-scale processing of human bone marrow cells. C 1992 Wiley-Liss, Inc.
Key words: autologous bone marrow transplantation, automatic processing of bone marrow cells, autotransplantation
INTRODUCTION
Autologous bone marrow transplantation (ABMT) permits the use of high doses of chemotherapy and radiotherapy in the treatment of cancer, since it is able to reverse the aplasia caused by these antineoplasic agents [ 11. Prior to performing in vivo or in vitro treatments, primitive hemapoietic progenitor cells are obtained from the bone marrow of the patient and cryopreserved [2-51. Engraftment speed, efficacy of in vitro treatments, and severity of side effects during and immediately following reinfusion are all dependent on the purity of the hematopoietic stem cells (SCs) collected [6]. Bone marrow concentration methods are geared to collect the greatest number of SCs possible so that hematopoietic function can be rapidly restored. These techniques also attempt to 1. Reduce the freezing volume in order to lessen the quantity of potentially toxic dimethylsulfoxide needed for freezing, as well as to save on storage space. 2. Eliminate the red blood cells (RBCs) that will be destroyed during the freezing-thawing process, thereby producing hemoglobinuria and hindering in vitro treatments. 0 1992 Wiley-Liss, Inc.
3. Remove nucleated cells, especially myeloid cells, that break down, tend to form aggregates, and cause SC loss. As in the case of RBCs, this also interferes with in vitro treatments [7,8]. Several techniques for the separation of bone marrow nucleated cells (NCs) have been employed, including sedimentation [9], buffy coat collection with blood cell separators [lo, 1 I], and mononuclear cell (MNC) isolation with blood cell processors 18,121. The latest blood cell separators allow the processing of SCs in a closed system, which simplifies separation procedures and improves the quality of the cells collected in terms of cell recoveries, collection selectivity, and microbiological safety. However, the blood cell processors use a semiautomated procedure, which requires Ficoll density gradients [ 121 or a prior centrifugation step to separate the hemo-
Received for publication November 1 1 , 1991; accepted September 1, 1992. Address reprint requests to Dr. Juan M. Rodriguez Fernandez, Jefe Servicio Hematologia-Hemoterapia, Hospital Universitario Virgen del Rocio, Avd. Manuel Siurot, SIN, 41013-Sevilla, Spain.
102
Rodriguez et al.
lysed plasma (personal experience). Tube connections are required by these methods and, although performed with sterile connecting devices under a laminar airflow hood, there is some risk of microbiological contamination. Furthermore, these methods are time-consuming . We report our experience in processing autologous bone marrow cell suspensions with a fully automated method using the CS-3000 blood cell separator (Fenwal Division, Baxter Healthcare Corp., Deerfield, IL). The laboratory data and clinical results after transplantation are compared to the data obtained with a semiautomated procedure using the same blood cell separator. The aims of this automated method are (1) to avoid the step necessary to eliminate the bone marrow fat and hemolysed plasma; (2) to eliminate the influence of the degree of operator technical skill on the final results; ( 3 ) to shorten the total working time significantly; (4) to maintain the high mononuclear cell recovery obtained using our semiautomated method; and (5) to reduce red cell and granulocyte contamination in a way which is highly reproducible.
PATIENTS AND METHODS Patients
Forty patients, ranging in age from 4 to 47 years and suffering from acute leukemia, lymphoma, or multiple myeloma, underwent bone marrow harvesting under general anesthesia. All bone marrow cell suspensions were processed using the automated CS-3000 blood cell separator. The bone marrow of 30 patients was processed by a semiautomated method, and a fully automated procedure was used for the remaining 10 cases. Of the 40 patients, 27 had undergone autologous bone marrow transplantation by the time of this report. The protocol was approved by the Investigation Committee of the Virgen del Rocio University Hospital. All risks of the procedures were fully explained to the patients and informed consent was obtained.
Bone Marrow Harvests
Bone marrow was harvested from patients’ iliac crests. The technique described by Thomas [2] was used in 9 cases and Fenwal collection/filtration kits were used on the remaining patients. Using the Fenwal kit, bone marrow is deposited in bags that can be sealed during the extraction procedure. The collected marrow is then strained through 3 metal filters (500, 300, and 200 microns, arranged in decreasing order) into a 600 or a 2,000 ml bag. At the end of this process, a 10% dilution of acid-citrate-dextrose, solution A (ACD-A), is added.
Bone Marrow Concentration
Procedures were performed on the Fenwal CS-3000 blood cell separator. This device is computerized and allows operational changes to be made to the program ~31.
Semiautomated Method
Thirty bone marrow cell suspensions were processed with a semiautomated method. This was the first bone marrow cell concentration method we used in our laboratory. The bag containing the filtered bone marrow was connected to a CS-3000 blood cell separator in a laminar airflow chamber. The central and lateral outputs of the bag were connected to the input and return lines of the separator, respectively. The bag was then placed on an agitator to homogenize the cellular contents, where it remained during the entire process. The procedure was carried out in 2 stages. Hemolyzed phase elimination stage. Plasma was extracted from the bone marrow bag following a modified version of the basic program for procedure number 1 of the CS3-000 (platelet procurement). The A-35 collection chamber and standard granulocyte separation chamber were used. The plasma volume to be collected from the bone marrow bag was calculated, taking into account the 450 ml of sodium chloride presents in the apheresis kit. The expected final bone marrow hematocrit is 85%. The interface was set at 100, the centrifuge speed at 1,600 rpm, and the blood flow rate at 50 ml/min. When the desired volume of plasma had been processed and the hemolysed plasma diverted into one of the PL-732 transfer pack units of the cell collection kit, the roller clamps of the inlet line, return line, and the tubing of the transfer pack units were closed; the halt irrigate switch was pressed, disconnecting both the inlet and return lines from the bone marrow bag. The concentration stage. In a laminar airflow chamber enough irradiated homologous or isogroup plasma was added to the bone marrow to achieve a final Hct of 25-30%. The bone marrow bag was then reconnected to the CS-3000. Procedure number 1 was of the modified CS-3000 program was employed. The interface was set at 200, the centrifuge speed at 1,400 rmp, and the blood flow rate at 50 ml/min. After the second spillover, the new baseline value was introduced into the computer. The interface was reduced to 150 whenever the pumped blood flowed between 25-35 ml/min. However, slower flows allow higher interferences (150-200) to be used. Once a volume equal to 5 times that of the bone marrow had been processed, procedure number 2 was instituted to extract the residue remaining in the separator. The bag was disconnected from the CS-3000, and MNC
-:.
Automated Mononuclear Bone Marrow Concentration
concentration in the A-35 chamber was agitated for a few minutes. Automated Method
Automated MNC isolation was performed on 10 collected bone marrow cell suspensions. The procedure, named 6/6 VP, is completely different from the standard CS-3000 programs. Preparation of the system and setting the external parameters. (a) A closed kit was used (#4R2230 Fenwal): 50 ml of a 20% solution of human albumin was perfused through a sample-site coupler attached to the bag containing a 0.9% sodium chloride solution. A granulocyte separation chamber and A-35 recollection chamber were used. The PRIME, RUN, and REINFUSE programs were entered into the computer’s memory in 1 of the 8 special programs of the cell separator (RUN-PROGRAM 6/6 VP). (b) External parameters: Centrifuge speed: 1,400 rpm Flow rate into the separator: 50 ml/min (blood flow rate) Interface detector: 50 (0500) Plasma collect in saline position: 200 ml END POINT: 2.5 X initial volume. Processing: PRIME, RUN, AND REINFUSE. (a) The kit was primed in the usual way (PRIME) using ACD-A as an anticoagulant. (b) RUN: The bag containing the bone marrow was connected to the inlet and return lines, and RUN was initiated using the 6/6 VP program parameters previously entered to separate the cells. With this kit the cell separator initially performs a control test on the first bone marrow fractionation in the granulocyte separation chamber, the first spillover taking place. Immediately before this, the baseline is entered, indicating the density of the hemolyzed plasma (Fig. 1). From this point, the 2 peristaltic pumps work to produce continuous fractions of the bone marrow through saturation of the separation chamber. The interface detector sensor controls the contamination of RBCs and granulocytes and reduces the speed of both pumps to 6 ml/min. Consequently, the first MNC fraction is confined to the section between the separation bags and the interface detector. Due to the low pump speed induced, the contents of the separation chamber sediment and a further fractionation of MNC occurs by means of centrifugal force. At the end of this stage, the blood pump attains a high flow rate (approximately 75 ml/min), expelling the previously sedimented red cells. At the same time the
103
PLASMA
1st STEP
P 2nd STEP
3rd STEP
4 t h STEP
Fig. 1. Automated procedure. Processing steps in the CS-3000: 1st step: The first fraction of MNCs is obtained. 2nd step: Detection of the first fraction of MNCs by the interface detector, first spillover. 3rd step: Formation of the second fraction of MNCs. 4th step: Both MNCs fraction are taken into the collection bag.
speed of the plasma pump also increases (approximately 25 ml/min), withdrawing the 2 MNC fractions produced during the spillover and depositing them in the recollection chamber. Once spillover has occurred, the blood pump returns to its initial speed of 50 ml/min and the plasma pump operates at the percentage of the speed of the blood pump dictated by the hematocrit value. The cycle is repeated as many times as spillovers occur during the process. (c) REINFUSE: Once the established END POINT is reached, the bone marrow remaining in the system is recuperated by means of REINFUSE. This procedure was carried out in a single stage. The same bag containing the marrow from the operating room is connected to the CS-3000 blood cell separator. Then, the program goes on until the MNC collection is completed without any need of manual assistance. Cryopreservation
Immediately after concentration, the MNCs suspended in autologous plasma were cryopreserved in a 10% dimethylsulfoxide solution in a programmed liquid nitrogen freezer (Cryoson BVlO) in standard fashion [ 14- 161. In Vitro Assaying
To study cell viability, samples were collected following filtration, concentration, and cryopreservation, and stained with a 0.04% solution of trypan blue. Directly thereafter, the specimens were cultured for CFUs-GM in a double layer of agar gel using a modified version of
104
Rodriguez et al.
the Pike and Robinson technique [17]. The buffy coat and filtered bone marrow cells were plated at a concentration of 1 x lo5. The MNCs were previously separated from the filtered bone marrow with a Ficoll-Hypaque density gradient. The slides were read “fresh” on an inverted microscope 14 days later and then after the cultures had been desiccated and stained. Groups of 50 or more cells were considered to be colonies. Complete cell counts of the bone marrow cell suspensions were performed using the Technicon H-1 System (Bayer). The number of MNCs recovered from the filtered and concentrated samples were evaluated by the cytocentrifugation method employing the following formula: Total cells x MNC percentage in concentrated Droduct Total cells X MNC percentage in filtered product
x 100. The progenitor cell recovery was calculated considering the number of processed and recovered CFU/GM colonies. Preparation for Transplant
Patients with acute nonlymphocytic leukemia (ANLL) received a 4-day course of busulfan (4 mg/kg/day) and then 2 days of cyclophosphamide (60 mg/kg/day). All except 2 of the acute lymphocytic leukemia (ALL) cases received total body irradiation delivered in one 950 Gy dose. The other 2 patients received eight 165 cGy doses, 1 every 12 hours during 4 consecutive days. Total body irradiation was followed by a 2-day course of cyclophosphamide (60 mg/kg). Patients with multiple myeloma also received this conditioning therapy. Hodgkin’s disease (HD) was treated with 1 day of BCNU (300 mg/ m2), 3 days of VP-16 (150 mg/m2), and 2 days of cyclophosphamide (60 mg/kg). Marrow was infused within 48 hours of the last dose of cyclophosphamide. Patients were hospitalized in single rooms with HEPA-filtered air. The end point for hematologic engraftment were defined as the time to engraftment, the number of days post-transplant until the granulocyte count reached >0.5 X 109/1and the untransfused platelet count reached >50 x 109/1. Statistical Analysis
Mann-Whitney test and the Pearson correlation test were used to assess differences between the results of both procedures and cloning efficiencies between processed and recovered MNCs.
RESULTS Laboratory Results
The parameters assessed were volume reduction; red blood cell contamination (measured as milliliters of RBC and hematocrit); granulocyte contamination; total number of NCs, MNCs, and CFU-GM; percentage of MNCs (lymphocytes and rnonocytes); and recovery of nucleated cells and MNCs in the final concentrate yielded by both the semiautomated and automated procedures. In addition, the time spent on both MNC isolation methods was taken into account. The final product processed by the blood cell separator in both procedures had a consistent volume of 200 ml, which corresponded to the capacity of the granulocyte separation chamber. Thus, volume is reduced by a greater percentage when a greater quantity of bone marrow is processed. The semiautomated MNC collection procedure allowed a total volume reduction of 83 ? 7% (range 6491). The concentrated product had an average red cell contamination of 21 ? 16 ml (range 4-73), with a mean hematocrit value of 11% (range 2-36) and a granulocyte contamination of 15% (range 1-42). The MNC percentage (lymphocytes and monocytes) in this final product was high, ranging from 35% to 96%, with a mean of 70 ? 16% (Tables I and 11). The NCs recovered ranged from 2.76 to 14.36 X lo9 (6.63 -C 2.29), and the recovery averaged 27.39% (range 14-46.7). The MNC recovery was consistently high and averaged 87.78 (range 57.1-98.7). The CFU-GM yield ranged from 42% to 34096, with an average of 97 ? 57%. Very high correlation was found between the cloning efficiencies of the processed and recovered MNCs ( P = 0.00001; Table I1 and Fig. 2). The entire semiautomated procedure took an average of 3 hours and 25 minutes to process a mean bone marrow suspension volume of 1,377 ml. The automated MNC isolation method took an average of 1 hour and 12 minutes to process a mean bone marrow suspension volume of 1,281 ml. An average of 27.11 7.16 NCs (range 12.62-37.35), with a mean MNC recovery of 86.98 7.07 (range 77-98.8%), were recovered in the final product. Cloning assays showed an average GFU-GM recovered of 98 2 60% (range 41-200). A good correlation was found between the number of CFU-GM assayed from the processed and recovered MNCs ( P = 0.01; Table III), which we believe will improve when the number of bone marrow cell suspensions processed increases (Fig. 3). The red cell contamination of the concentrated product was very low [9 ml; hematocrit 5% and very reproducible, ranging from 6 to 12 ml of red blood cells and 3 to 6% hematocrit (Table I11 and IV)], and the granulocyte contamination
*
*
Automated Mononuclear Bone Marrow Concentration TABLE I. Semiautomated Method Results (30 Cases)" Case
~
I 2 3 4 5 6 7 8 9 10
I1 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Mean Range
Granulocytes (%)
Hematocnt (%) ~~
~
7 4 20 17 8 12 9 15 9 17 I 4 20 ND 8 8 ND 15 39 15 18 30 23 7 29 2 24 II 42 11
5 8 17 3 13 10 7 7 9 8 2 5 17 36 4 8
ND 5 24 10 11
II 5 17 36 7 8 7 9 6 11 ? 8 2-36
15
k 10 1-42
MNC (%) ~~
93 96 69 80 89 72 85 64 56 67 84 90 69 50 90 60 ND 85 64 48 74 62 51 67 83 35 85 66 43 63 7 0 ? 16 35-96
*Bone marrow cell concentration using the semiautomated procedure. Red cell and granulocyte contamination. and MNC percentage in the final product.
averaged 18% (range 11-31), with average MNCs (lym12%, range 45-84 phocytes and monocytes) of 67 (Table IV). When both procedures were compared (MannWhitney test, Tables I1 and 111), statistically significant differences in NC, MNC recoveries, and MNC cloning efficiencies (CFU-GM) were not found. The mean values of the red cell contamination were significantly different, being lower and more reproducible in the automated procedure. Furthermore, the time spent to process 1 liter of bone marrow cell suspensions is considerably shorter in the automated method.
*
Clinical Results
Twenty-seven patients had been transplanted by the time this paper was written. Of the 27 marrows, 22 were processed by the semiautomated method and 5 were processed by the automatic procedure. All 27 patients experienced a complete hematologic recovery. The times to
105
engraftment (in days) for neutrophils and platelets are shown for all patients in Tables V and VI. Median time to engraftment of neutrophils to >500 X 106/1was 21 days for the semiautomatic-processed marrows and 20 days for the automatic-concentrated marrows, and the median times to engraftment of platelets to an untransfused level of >50 X 109/1were 39 days and 38 days, respectively. DISCUSSION
The main objective of bone marrow concentration techniques is to recover the maximum number of MNCs possible, since they are considered to be a source of hematopoietic progenitor cells which are capable of completely repopulating and maintaining bone marrow. An exact procedure to identify these hematopoietic SCs is still not available. Therefore, most authors study the ability of lymphocytes and monocytes to develop colonies in vitro, even though the number of MNCs obtained does not correlate with CFU-GM growth. All the bone marrow concentration methods aim to reduce the infusion volume as much as possible in order to lower the quantity of dimethylsulfoxide needed in the freezing process. The adverse side effects can thereby be avoided, and the cost of reagents necessary for in vitro treatment can be reduced. Likewise, postinfusion hemoglobinuria can be avoided by reducing the number of RBCs and Hct values adjusted to optimum levels for in vitro treatment. Nucleated cells, especially myeloid, and RBCs interfere with the activity of the cyclophosphamide derivatives 4-HC and ASTZ-Z, as well as with the activity of monoclonal antibody and complement [7,18] when the bone marrow is purged in vitro. Thus, to attain uniform cytolytic effects and to save on drugs, a reduction of the hematocrit value to 7-10% [7], or to below 2% [5], as well as a low granulocyte content [ 181, is needed when performing treatment either with the cyclophosphamide derivatives, 4-HC and ASTA-Z, or monoclonal antibodies plus complement. Therefore, a bone marrow concentration technique should be able to reduce the MNC final volume as much as possible, as well as its granulocyte and RBC content, and to recover the maximum number of the MNCs of the harvested bone marrow to retain marrow viability. On the other hand, it should be automated, performed in a closed, sterile system, totally reproducible, and time efficient. Forty autologous bone marrow harvests were processed using the CS-3000 cell processor. Two bone marrow concentration methods were utilized. The first of these was a semiautomated technique that, before MNC concentration, separates out most of the hemolysed plasma, culture medium, and the fat left after filtration,
Rodriguez et al.
106
TABLE 11. Semiautomated Method Results (30 Cases)*
Total volume Mean Range Red cell volume Mean Range
NCsb.' Mean Range MNCS~.~ Mean Range
Input (ml)
Recovered (ml)
Removal (%)
1.300 t 425 561-2255
200 (constant)
83 t 7 64.34-1.13
401 t 160 171-803
21 t 16 4-73
94 ? 4 80-98
Input ( x lo9)
Recovered ( X 10')
Recovery (%)
24.80 t 6.88 13.83-43.97
6.63 t 2.29 2.76- 14.36
27.39 -C 8.36 14.03-46.73
5.14 t 2.22 1.30- 14.36
4.47 t 1.91 1.18-12.72
87.78 -C 11.48 57.14-98.7
Input ( x lo6)
CFU-GM~.~ Mean Range
5.23 t 5a 0.8-22
Recovered
(X
4.44 t 4a 0.8-16
lo6)
Recovery (70) 97 t 57 42-340
*Total volume, red cell volume, NCs and MNCs recovery, and cloning efficiency (CFUs-GM) of bone marrow cell before and after the CS-3000 semiautomated concentration procedure. 'Pearson correlation coefficient = 0.736 ( P = 0.00001). bMann-Whitney test. 'NCs = nucleated cells. d~~~~ = mononuclear cells. 'CFU-GM = granulocyte-macrophage colony-forming unit.
1
'
~
1
"
1
"
1
'
'
"
24
1 0 w
'
1
~
'
i
r=0.79 p
A Fully Automated Method for Mononuclear Bone Marrow Cell Concentration J.M. Rodriguez, M. Carmona, P. Noguerol, M. Ruiz, R. Parody, J.M. Perez-Hurtado, and I. Espigado
F. Vidal,
Department of Hematology, Virgen del Rocio University Hospital, Seville, Spain (J.M. R., M.C., P.N., M.R., R.P., J.M.P.-H., I.E.) and Baxter, Spain (F.V.) We describe our experience in processing 40 bone marrow aspirates harvested for autotransplantation from patients with several hematological diseases using the CS-3000 blood cell separator. The bone marrow of the first 30 patients was processed by a semiautomated method, and a fully automated procedure was used for the remaining 10 cases. Both procedures were developed in our laboratory and yielded a similar average mononuclear cell recovery of 87.78% and 86.988, respectively, and similar nucleated cell recovery (27.39% and 27.11%). The cloning efficiency of hematopoietic progenitor cells, measured as the total CFU-GM colony recovery in the in vitro cultures, did not differ between processed and recovered mononuclear cells. On the other hand, all the patients with transplants showed complete hematologic recovery, and the time to engraftment was similar to that described for other procedures. The automated procedure resulted in an average red cell removal of 97.81%, similar to the semiautomated procedure (94.198), though with a narrower range (96.31-98.6% vs. 80.34-98.34%). The time taken to process a similar amount of bone marrow cell suspension was very different for each method: 1 hour for the fully automated vs. 2’12 hours for the semiautomated method to process 1,000 ml. Furthermore, the semiautomated procedure required the addition of homologous or irradiated plasma in a laminar air flow chamber, while the automated method is performed in a closed sterile system. We conclude that our procedure using the CS-3000 processor is an efficient method for fully automated large-scale processing of human bone marrow cells. C 1992 Wiley-Liss, Inc.
Key words: autologous bone marrow transplantation, automatic processing of bone marrow cells, autotransplantation
INTRODUCTION
Autologous bone marrow transplantation (ABMT) permits the use of high doses of chemotherapy and radiotherapy in the treatment of cancer, since it is able to reverse the aplasia caused by these antineoplasic agents [ 11. Prior to performing in vivo or in vitro treatments, primitive hemapoietic progenitor cells are obtained from the bone marrow of the patient and cryopreserved [2-51. Engraftment speed, efficacy of in vitro treatments, and severity of side effects during and immediately following reinfusion are all dependent on the purity of the hematopoietic stem cells (SCs) collected [6]. Bone marrow concentration methods are geared to collect the greatest number of SCs possible so that hematopoietic function can be rapidly restored. These techniques also attempt to 1. Reduce the freezing volume in order to lessen the quantity of potentially toxic dimethylsulfoxide needed for freezing, as well as to save on storage space. 2. Eliminate the red blood cells (RBCs) that will be destroyed during the freezing-thawing process, thereby producing hemoglobinuria and hindering in vitro treatments. 0 1992 Wiley-Liss, Inc.
3. Remove nucleated cells, especially myeloid cells, that break down, tend to form aggregates, and cause SC loss. As in the case of RBCs, this also interferes with in vitro treatments [7,8]. Several techniques for the separation of bone marrow nucleated cells (NCs) have been employed, including sedimentation [9], buffy coat collection with blood cell separators [lo, 1 I], and mononuclear cell (MNC) isolation with blood cell processors 18,121. The latest blood cell separators allow the processing of SCs in a closed system, which simplifies separation procedures and improves the quality of the cells collected in terms of cell recoveries, collection selectivity, and microbiological safety. However, the blood cell processors use a semiautomated procedure, which requires Ficoll density gradients [ 121 or a prior centrifugation step to separate the hemo-
Received for publication November 1 1 , 1991; accepted September 1, 1992. Address reprint requests to Dr. Juan M. Rodriguez Fernandez, Jefe Servicio Hematologia-Hemoterapia, Hospital Universitario Virgen del Rocio, Avd. Manuel Siurot, SIN, 41013-Sevilla, Spain.
102
Rodriguez et al.
lysed plasma (personal experience). Tube connections are required by these methods and, although performed with sterile connecting devices under a laminar airflow hood, there is some risk of microbiological contamination. Furthermore, these methods are time-consuming . We report our experience in processing autologous bone marrow cell suspensions with a fully automated method using the CS-3000 blood cell separator (Fenwal Division, Baxter Healthcare Corp., Deerfield, IL). The laboratory data and clinical results after transplantation are compared to the data obtained with a semiautomated procedure using the same blood cell separator. The aims of this automated method are (1) to avoid the step necessary to eliminate the bone marrow fat and hemolysed plasma; (2) to eliminate the influence of the degree of operator technical skill on the final results; ( 3 ) to shorten the total working time significantly; (4) to maintain the high mononuclear cell recovery obtained using our semiautomated method; and (5) to reduce red cell and granulocyte contamination in a way which is highly reproducible.
PATIENTS AND METHODS Patients
Forty patients, ranging in age from 4 to 47 years and suffering from acute leukemia, lymphoma, or multiple myeloma, underwent bone marrow harvesting under general anesthesia. All bone marrow cell suspensions were processed using the automated CS-3000 blood cell separator. The bone marrow of 30 patients was processed by a semiautomated method, and a fully automated procedure was used for the remaining 10 cases. Of the 40 patients, 27 had undergone autologous bone marrow transplantation by the time of this report. The protocol was approved by the Investigation Committee of the Virgen del Rocio University Hospital. All risks of the procedures were fully explained to the patients and informed consent was obtained.
Bone Marrow Harvests
Bone marrow was harvested from patients’ iliac crests. The technique described by Thomas [2] was used in 9 cases and Fenwal collection/filtration kits were used on the remaining patients. Using the Fenwal kit, bone marrow is deposited in bags that can be sealed during the extraction procedure. The collected marrow is then strained through 3 metal filters (500, 300, and 200 microns, arranged in decreasing order) into a 600 or a 2,000 ml bag. At the end of this process, a 10% dilution of acid-citrate-dextrose, solution A (ACD-A), is added.
Bone Marrow Concentration
Procedures were performed on the Fenwal CS-3000 blood cell separator. This device is computerized and allows operational changes to be made to the program ~31.
Semiautomated Method
Thirty bone marrow cell suspensions were processed with a semiautomated method. This was the first bone marrow cell concentration method we used in our laboratory. The bag containing the filtered bone marrow was connected to a CS-3000 blood cell separator in a laminar airflow chamber. The central and lateral outputs of the bag were connected to the input and return lines of the separator, respectively. The bag was then placed on an agitator to homogenize the cellular contents, where it remained during the entire process. The procedure was carried out in 2 stages. Hemolyzed phase elimination stage. Plasma was extracted from the bone marrow bag following a modified version of the basic program for procedure number 1 of the CS3-000 (platelet procurement). The A-35 collection chamber and standard granulocyte separation chamber were used. The plasma volume to be collected from the bone marrow bag was calculated, taking into account the 450 ml of sodium chloride presents in the apheresis kit. The expected final bone marrow hematocrit is 85%. The interface was set at 100, the centrifuge speed at 1,600 rpm, and the blood flow rate at 50 ml/min. When the desired volume of plasma had been processed and the hemolysed plasma diverted into one of the PL-732 transfer pack units of the cell collection kit, the roller clamps of the inlet line, return line, and the tubing of the transfer pack units were closed; the halt irrigate switch was pressed, disconnecting both the inlet and return lines from the bone marrow bag. The concentration stage. In a laminar airflow chamber enough irradiated homologous or isogroup plasma was added to the bone marrow to achieve a final Hct of 25-30%. The bone marrow bag was then reconnected to the CS-3000. Procedure number 1 was of the modified CS-3000 program was employed. The interface was set at 200, the centrifuge speed at 1,400 rmp, and the blood flow rate at 50 ml/min. After the second spillover, the new baseline value was introduced into the computer. The interface was reduced to 150 whenever the pumped blood flowed between 25-35 ml/min. However, slower flows allow higher interferences (150-200) to be used. Once a volume equal to 5 times that of the bone marrow had been processed, procedure number 2 was instituted to extract the residue remaining in the separator. The bag was disconnected from the CS-3000, and MNC
-:.
Automated Mononuclear Bone Marrow Concentration
concentration in the A-35 chamber was agitated for a few minutes. Automated Method
Automated MNC isolation was performed on 10 collected bone marrow cell suspensions. The procedure, named 6/6 VP, is completely different from the standard CS-3000 programs. Preparation of the system and setting the external parameters. (a) A closed kit was used (#4R2230 Fenwal): 50 ml of a 20% solution of human albumin was perfused through a sample-site coupler attached to the bag containing a 0.9% sodium chloride solution. A granulocyte separation chamber and A-35 recollection chamber were used. The PRIME, RUN, and REINFUSE programs were entered into the computer’s memory in 1 of the 8 special programs of the cell separator (RUN-PROGRAM 6/6 VP). (b) External parameters: Centrifuge speed: 1,400 rpm Flow rate into the separator: 50 ml/min (blood flow rate) Interface detector: 50 (0500) Plasma collect in saline position: 200 ml END POINT: 2.5 X initial volume. Processing: PRIME, RUN, AND REINFUSE. (a) The kit was primed in the usual way (PRIME) using ACD-A as an anticoagulant. (b) RUN: The bag containing the bone marrow was connected to the inlet and return lines, and RUN was initiated using the 6/6 VP program parameters previously entered to separate the cells. With this kit the cell separator initially performs a control test on the first bone marrow fractionation in the granulocyte separation chamber, the first spillover taking place. Immediately before this, the baseline is entered, indicating the density of the hemolyzed plasma (Fig. 1). From this point, the 2 peristaltic pumps work to produce continuous fractions of the bone marrow through saturation of the separation chamber. The interface detector sensor controls the contamination of RBCs and granulocytes and reduces the speed of both pumps to 6 ml/min. Consequently, the first MNC fraction is confined to the section between the separation bags and the interface detector. Due to the low pump speed induced, the contents of the separation chamber sediment and a further fractionation of MNC occurs by means of centrifugal force. At the end of this stage, the blood pump attains a high flow rate (approximately 75 ml/min), expelling the previously sedimented red cells. At the same time the
103
PLASMA
1st STEP
P 2nd STEP
3rd STEP
4 t h STEP
Fig. 1. Automated procedure. Processing steps in the CS-3000: 1st step: The first fraction of MNCs is obtained. 2nd step: Detection of the first fraction of MNCs by the interface detector, first spillover. 3rd step: Formation of the second fraction of MNCs. 4th step: Both MNCs fraction are taken into the collection bag.
speed of the plasma pump also increases (approximately 25 ml/min), withdrawing the 2 MNC fractions produced during the spillover and depositing them in the recollection chamber. Once spillover has occurred, the blood pump returns to its initial speed of 50 ml/min and the plasma pump operates at the percentage of the speed of the blood pump dictated by the hematocrit value. The cycle is repeated as many times as spillovers occur during the process. (c) REINFUSE: Once the established END POINT is reached, the bone marrow remaining in the system is recuperated by means of REINFUSE. This procedure was carried out in a single stage. The same bag containing the marrow from the operating room is connected to the CS-3000 blood cell separator. Then, the program goes on until the MNC collection is completed without any need of manual assistance. Cryopreservation
Immediately after concentration, the MNCs suspended in autologous plasma were cryopreserved in a 10% dimethylsulfoxide solution in a programmed liquid nitrogen freezer (Cryoson BVlO) in standard fashion [ 14- 161. In Vitro Assaying
To study cell viability, samples were collected following filtration, concentration, and cryopreservation, and stained with a 0.04% solution of trypan blue. Directly thereafter, the specimens were cultured for CFUs-GM in a double layer of agar gel using a modified version of
104
Rodriguez et al.
the Pike and Robinson technique [17]. The buffy coat and filtered bone marrow cells were plated at a concentration of 1 x lo5. The MNCs were previously separated from the filtered bone marrow with a Ficoll-Hypaque density gradient. The slides were read “fresh” on an inverted microscope 14 days later and then after the cultures had been desiccated and stained. Groups of 50 or more cells were considered to be colonies. Complete cell counts of the bone marrow cell suspensions were performed using the Technicon H-1 System (Bayer). The number of MNCs recovered from the filtered and concentrated samples were evaluated by the cytocentrifugation method employing the following formula: Total cells x MNC percentage in concentrated Droduct Total cells X MNC percentage in filtered product
x 100. The progenitor cell recovery was calculated considering the number of processed and recovered CFU/GM colonies. Preparation for Transplant
Patients with acute nonlymphocytic leukemia (ANLL) received a 4-day course of busulfan (4 mg/kg/day) and then 2 days of cyclophosphamide (60 mg/kg/day). All except 2 of the acute lymphocytic leukemia (ALL) cases received total body irradiation delivered in one 950 Gy dose. The other 2 patients received eight 165 cGy doses, 1 every 12 hours during 4 consecutive days. Total body irradiation was followed by a 2-day course of cyclophosphamide (60 mg/kg). Patients with multiple myeloma also received this conditioning therapy. Hodgkin’s disease (HD) was treated with 1 day of BCNU (300 mg/ m2), 3 days of VP-16 (150 mg/m2), and 2 days of cyclophosphamide (60 mg/kg). Marrow was infused within 48 hours of the last dose of cyclophosphamide. Patients were hospitalized in single rooms with HEPA-filtered air. The end point for hematologic engraftment were defined as the time to engraftment, the number of days post-transplant until the granulocyte count reached >0.5 X 109/1and the untransfused platelet count reached >50 x 109/1. Statistical Analysis
Mann-Whitney test and the Pearson correlation test were used to assess differences between the results of both procedures and cloning efficiencies between processed and recovered MNCs.
RESULTS Laboratory Results
The parameters assessed were volume reduction; red blood cell contamination (measured as milliliters of RBC and hematocrit); granulocyte contamination; total number of NCs, MNCs, and CFU-GM; percentage of MNCs (lymphocytes and rnonocytes); and recovery of nucleated cells and MNCs in the final concentrate yielded by both the semiautomated and automated procedures. In addition, the time spent on both MNC isolation methods was taken into account. The final product processed by the blood cell separator in both procedures had a consistent volume of 200 ml, which corresponded to the capacity of the granulocyte separation chamber. Thus, volume is reduced by a greater percentage when a greater quantity of bone marrow is processed. The semiautomated MNC collection procedure allowed a total volume reduction of 83 ? 7% (range 6491). The concentrated product had an average red cell contamination of 21 ? 16 ml (range 4-73), with a mean hematocrit value of 11% (range 2-36) and a granulocyte contamination of 15% (range 1-42). The MNC percentage (lymphocytes and monocytes) in this final product was high, ranging from 35% to 96%, with a mean of 70 ? 16% (Tables I and 11). The NCs recovered ranged from 2.76 to 14.36 X lo9 (6.63 -C 2.29), and the recovery averaged 27.39% (range 14-46.7). The MNC recovery was consistently high and averaged 87.78 (range 57.1-98.7). The CFU-GM yield ranged from 42% to 34096, with an average of 97 ? 57%. Very high correlation was found between the cloning efficiencies of the processed and recovered MNCs ( P = 0.00001; Table I1 and Fig. 2). The entire semiautomated procedure took an average of 3 hours and 25 minutes to process a mean bone marrow suspension volume of 1,377 ml. The automated MNC isolation method took an average of 1 hour and 12 minutes to process a mean bone marrow suspension volume of 1,281 ml. An average of 27.11 7.16 NCs (range 12.62-37.35), with a mean MNC recovery of 86.98 7.07 (range 77-98.8%), were recovered in the final product. Cloning assays showed an average GFU-GM recovered of 98 2 60% (range 41-200). A good correlation was found between the number of CFU-GM assayed from the processed and recovered MNCs ( P = 0.01; Table III), which we believe will improve when the number of bone marrow cell suspensions processed increases (Fig. 3). The red cell contamination of the concentrated product was very low [9 ml; hematocrit 5% and very reproducible, ranging from 6 to 12 ml of red blood cells and 3 to 6% hematocrit (Table I11 and IV)], and the granulocyte contamination
*
*
Automated Mononuclear Bone Marrow Concentration TABLE I. Semiautomated Method Results (30 Cases)" Case
~
I 2 3 4 5 6 7 8 9 10
I1 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Mean Range
Granulocytes (%)
Hematocnt (%) ~~
~
7 4 20 17 8 12 9 15 9 17 I 4 20 ND 8 8 ND 15 39 15 18 30 23 7 29 2 24 II 42 11
5 8 17 3 13 10 7 7 9 8 2 5 17 36 4 8
ND 5 24 10 11
II 5 17 36 7 8 7 9 6 11 ? 8 2-36
15
k 10 1-42
MNC (%) ~~
93 96 69 80 89 72 85 64 56 67 84 90 69 50 90 60 ND 85 64 48 74 62 51 67 83 35 85 66 43 63 7 0 ? 16 35-96
*Bone marrow cell concentration using the semiautomated procedure. Red cell and granulocyte contamination. and MNC percentage in the final product.
averaged 18% (range 11-31), with average MNCs (lym12%, range 45-84 phocytes and monocytes) of 67 (Table IV). When both procedures were compared (MannWhitney test, Tables I1 and 111), statistically significant differences in NC, MNC recoveries, and MNC cloning efficiencies (CFU-GM) were not found. The mean values of the red cell contamination were significantly different, being lower and more reproducible in the automated procedure. Furthermore, the time spent to process 1 liter of bone marrow cell suspensions is considerably shorter in the automated method.
*
Clinical Results
Twenty-seven patients had been transplanted by the time this paper was written. Of the 27 marrows, 22 were processed by the semiautomated method and 5 were processed by the automatic procedure. All 27 patients experienced a complete hematologic recovery. The times to
105
engraftment (in days) for neutrophils and platelets are shown for all patients in Tables V and VI. Median time to engraftment of neutrophils to >500 X 106/1was 21 days for the semiautomatic-processed marrows and 20 days for the automatic-concentrated marrows, and the median times to engraftment of platelets to an untransfused level of >50 X 109/1were 39 days and 38 days, respectively. DISCUSSION
The main objective of bone marrow concentration techniques is to recover the maximum number of MNCs possible, since they are considered to be a source of hematopoietic progenitor cells which are capable of completely repopulating and maintaining bone marrow. An exact procedure to identify these hematopoietic SCs is still not available. Therefore, most authors study the ability of lymphocytes and monocytes to develop colonies in vitro, even though the number of MNCs obtained does not correlate with CFU-GM growth. All the bone marrow concentration methods aim to reduce the infusion volume as much as possible in order to lower the quantity of dimethylsulfoxide needed in the freezing process. The adverse side effects can thereby be avoided, and the cost of reagents necessary for in vitro treatment can be reduced. Likewise, postinfusion hemoglobinuria can be avoided by reducing the number of RBCs and Hct values adjusted to optimum levels for in vitro treatment. Nucleated cells, especially myeloid, and RBCs interfere with the activity of the cyclophosphamide derivatives 4-HC and ASTZ-Z, as well as with the activity of monoclonal antibody and complement [7,18] when the bone marrow is purged in vitro. Thus, to attain uniform cytolytic effects and to save on drugs, a reduction of the hematocrit value to 7-10% [7], or to below 2% [5], as well as a low granulocyte content [ 181, is needed when performing treatment either with the cyclophosphamide derivatives, 4-HC and ASTA-Z, or monoclonal antibodies plus complement. Therefore, a bone marrow concentration technique should be able to reduce the MNC final volume as much as possible, as well as its granulocyte and RBC content, and to recover the maximum number of the MNCs of the harvested bone marrow to retain marrow viability. On the other hand, it should be automated, performed in a closed, sterile system, totally reproducible, and time efficient. Forty autologous bone marrow harvests were processed using the CS-3000 cell processor. Two bone marrow concentration methods were utilized. The first of these was a semiautomated technique that, before MNC concentration, separates out most of the hemolysed plasma, culture medium, and the fat left after filtration,
Rodriguez et al.
106
TABLE 11. Semiautomated Method Results (30 Cases)*
Total volume Mean Range Red cell volume Mean Range
NCsb.' Mean Range MNCS~.~ Mean Range
Input (ml)
Recovered (ml)
Removal (%)
1.300 t 425 561-2255
200 (constant)
83 t 7 64.34-1.13
401 t 160 171-803
21 t 16 4-73
94 ? 4 80-98
Input ( x lo9)
Recovered ( X 10')
Recovery (%)
24.80 t 6.88 13.83-43.97
6.63 t 2.29 2.76- 14.36
27.39 -C 8.36 14.03-46.73
5.14 t 2.22 1.30- 14.36
4.47 t 1.91 1.18-12.72
87.78 -C 11.48 57.14-98.7
Input ( x lo6)
CFU-GM~.~ Mean Range
5.23 t 5a 0.8-22
Recovered
(X
4.44 t 4a 0.8-16
lo6)
Recovery (70) 97 t 57 42-340
*Total volume, red cell volume, NCs and MNCs recovery, and cloning efficiency (CFUs-GM) of bone marrow cell before and after the CS-3000 semiautomated concentration procedure. 'Pearson correlation coefficient = 0.736 ( P = 0.00001). bMann-Whitney test. 'NCs = nucleated cells. d~~~~ = mononuclear cells. 'CFU-GM = granulocyte-macrophage colony-forming unit.
1
'
~
1
"
1
"
1
'
'
"
24
1 0 w
'
1
~
'
i
r=0.79 p
