TRANSPLANTATION AND CELLULAR ENGINEERING Red blood cell depletion from bone marrow and peripheral blood buffy coat: a comparison of two new and three established technologies € mmer,1 Nadine Sorg,1* Carolin Poppe,1* Milica Bunos,1 Eva Wingenfeld,1 Christiane Hu € mer,1 Belinda Stock,2 Erhard Seifried,1,2 and Halvard Bonig1,2,3 Ariane Kra
BACKGROUND: Red blood cell (RBC) depletion is a standard technique for preparation of ABO-incompatible bone marrow transplants (BMTs). Density centrifugation or apheresis are used successfully at clinical scale. The advent of a bone marrow (BM) processing module for the Spectra Optia (Terumo BCT) provided the initiative to formally compare our standard technology, the COBE2991 (Ficoll, manual, “C”) with the Spectra Optia BMP (apheresis, semiautomatic, “O”), the Sepax II NeatCell (Ficoll, automatic, “S”), the Miltenyi CliniMACS Prodigy density gradient separation system (Ficoll, automatic, “P”), and manual Ficoll (“M”). C and O handle larger product volumes than S, P, and M. STUDY DESIGN AND METHODS: Technologies were assessed for RBC depletion, target cell (mononuclear cells [MNCs] for buffy coats [BCs], CD341 cells for BM) recovery, and cost/labor. BC pools were simultaneously purged with C, O, S, and P; five to 18 BM samples were sequentially processed with C, O, S, and M. RESULTS: Mean RBC removal with C was 97% (BCs) or 92% (BM). From both products, O removed 97%, and P, S, and M removed 99% of RBCs. MNC recovery from BC (98% C, 97% O, 65% P, 74% S) or CD341 cell recovery from BM (92% C, 90% O, 67% S, 70% M) were best with C and O. Polymorphonuclear cells (PMNs) were depleted from BCs by P, S, and C, while O recovered 50% of PMNs. Time savings compared to C or M for all tested technologies are considerable. CONCLUSION: All methods are in principle suitable and can be selected based on sample volume, available technology, and desired product specifications beyond RBC depletion and MNC and/or CD341 cell recovery.
S
ince HLA and A-/B-transferase genes are inherited independently, an ABO major or bidirectional mismatch is observed in roughly 20% of bone marrow transplantations (BMTs),1 that is, approximately 600 times per year in all of Europe with its more than 600 transplant programs2 or once per year per center. A typical dose of bone marrow (BM) cells for BMT is 2.5 3 108 to 3.0 3 108/kg nucleated cells,3-5 equivalent to approximately 12 to 20 mL of BM aspirate/kg. At a mean hematocrit (Hct) of 30% to 35%, the red blood cell (RBC) content of BM aspirate is thus considerable, 4 to 7 mL/kg of the recipient, and not unconditionally tolerable for ABO major–mismatched recipients. One long-established method to avoid severe and potentially lethal hemolysis ABBREVIATIONS: BC 5 buffy coat; BM 5 bone marrow. From the 1German Red Cross Blood Service Baden€rttemberg-Hessen, Institute Frankfurt, and the 2Institute for Wu Transfusion Medicine and Immunohematology, Goethe University, Frankfurt, Germany; and the 3Department of Medicine/Hematology, University of Washington, Seattle, Washington. € nig, Goethe University Address reprint requests to: Halvard Bo Medical Center, Institute for Transfusion Medicine and Immunohematology, and German Red Cross Blood Service Baden€rttemberg-Hesse, Sandhofstrasse 1, 60528 Frankfurt, Germany; Wu e-mail: [email protected] [email protected]. *Contributed equally. HB and ES are members of the LOEWE Cell and Gene Therapy Frankfurt faculty funded by Hessian Ministry of Higher Education, Research and the Arts ref.no.: III L 4-518/ 17.004 (2010). RBC-depletion tubing sets for Prodigy were gifted by Miltenyi Biotech. No outside funding was used to support the studies. Received for publication September 24, 2014; revision received November 24, 2014; and accepted November 28, 2014. doi:10.1111/trf.13001 C 2015 AABB V
TRANSFUSION 2015;55;1275–1282 Volume 55, June 2015 TRANSFUSION 1275
SORG ET AL.
during BM infusion is RBC depletion from the graft.6 RBC depletion is also used for volume reduction of BM aspirate before cryoconservation. Since RBCs, white blood cells (WBCs), platelets (PLTs), and plasma differ in size and density, they can be efficiently separated by density gradient centrifugation or apheresis, and both have been applied successfully.6-9 The critical variables for the success of RBC depletion of BM are recovery of CD341 “stem cells,” which are contained within the mononuclear cell (MNC) population and, when performed in the context of ABO major– incompatible BMT, near-qualitative removal of RBCs to values below 0.5 mL/kg body weight of the recipient. Similar processing of whole blood, buffy coats (BCs), and occasionally, apheresis products generally also aims for best-possible RBC depletion while maximally recovering MNCs. Density gradient centrifugation with “Ficoll,” a synthetic low-viscosity, high-molecular-weight hydrophilic sucrose, and epichlorohydrine polymer and sodium diatrizoate, has been extensively used for RBC (and PMN) depletion.10 Alternatively, apheresis technology has been applied to deplete RBCs from BM. The modest differences in separation physics between density gradient centrifugation and apheresis allow some prediction of RBC depletion outcomes. In terms of available technology, the COBE29916,11-13 (Terumo BCT, Lakewood, CO) and the Sepax II NeatCell14,15 (Biosafe SA, Eysins, Switzerland) have been on the market for many decades and several years, respectively. Both are based on Ficoll density centrifugation, and significant evidence of their suitability has been provided. The Spectra Optia BMP is a new application for the relatively recent Spectra Optia apheresis device that was previously introduced.16-24 The package consists of the standard Spectra Optia MNC filler, the standard Spectra Optia IDL tubing set, and a large BM bag for recirculating of the (increasingly MNC-deplete) BM suspension, as well as a software that provides fully automatic, photo detector– guided apheresis of the BM, physically highly analogous to MNC apheresis in stem cell donors. The Spectra Optia BMP is licensed in the European Medicines Agency region, but not in the United States. The CliniMACS Prodigy (Miltenyi Biotec, Bergisch Gladbach, Germany) device is a multifunctional automatic cell processing device consisting of, among other features, a camera-fitted centrifugation chamber, where the device can locate, track, and selectively aspirate the MNC layer; a roller pump; valves; and a single-use sterile all-in-one consumable.25 The Ficoll package for Prodigy has not yet been publically released; a near-series beta version of the software and prototypic plasticware provided by the manufacturer were used for the validation. For neither of the two latter methods have performance data been reported. In a side-byside validation, large BC pools from 20 donors each were 1276 TRANSFUSION Volume 55, June 2015
split and subjected to concurrent RBC depletion with all four technologies. Moreover, 53 clinical BM aspirates were RBC depleted with the COBE2991, the Spectra Optia BMP, the Sepax II NeatCell, or manual Ficoll. Results indicating feasibility and clinically adequate efficiency, albeit with some quantitative variation, are presented.
MATERIALS AND METHODS BC pools were generated from residual BCs from 20 healthy, ABO-identical whole blood donors with donor and internal review board approval (permit number 329/ 19). Two large (approx. 500 mL) and two small (100 mL) aliquots were concurrently separated with four different technologies. BM samples were clinical BM aspirates for hematopoietic reconstitution (COBE2991, n 5 14; Spectra Optia, n 5 5), mesenchymal stroma cell isolation (Sepax, n 5 16), or cardiovascular regeneration (manual Ficoll, n 5 18); routine performance data were collected after pseudonymization. Product quality including Hct was assessed by automatic hemocytometry (Sysmex XT-1800i hematology analyzer, Sysmex, Norderstedt, Germany); in the processed marrow samples, CD341 cells were enumerated using the single-platform stem cell enumeration kit and flow cytometer (FACSCalibur, Becton-Dickinson, Heidelberg, Germany), as described.26 Volume was assessed by weight with correction for Hct. Reagents required for the RBC depletions include density medium (Ficoll, Lonza, Verviers, Belgium), as well as NaCl 0.9% and human serum albumin (Baxter Healthcare, Unterschleißheim, Germany).
Separation devices and protocols One manual (COBE2991), one semiautomatic (Spectra Optia) and two fully automatic, walk-away (Sepax II and CliniMACS Prodigy) RBC depletion protocols were used, in addition to fully manual semiopen Ficoll processing in 50-mL conical tubes. The COBE2991 device was fitted with the corresponding consumable (Terumo BCT); the manual process was performed according to standard operating procedures based on the manufacturer’s instructions. Briefly, separation of WBCs and RBCs was achieved by layering Ficoll under the BC in the centrifugation bag (donut bag) and extraction of the different phases by inflating the cushion underneath and manual clamping or unclamping of the waste, nontarget, and target bags based on visual detection of the different fractions in the tubing set. The Spectra Optia, previously described with respect to its performance as patient or donor apheresis device,16-24 was fitted with the standard filler (as for MNC apheresis), the IDL separation kit, and the BMP accessory kit. The set was installed and primed as directed by the manufacturer. An initial collection preference16 of 50 was selected, as manufacturer recommended; RBC depletions
NEW RBC DEPLETION TECHNOLOGIES
TABLE 1. Performance data of different RBC depletion technologies with BC pools RBC depletion technology Before processing (volume, mL) Preprocess Hct (%) Preprocess RBC volume (mL) Preprocess MNC count (3109/L) PLT count (3109/L) After processing (volume, mL) Postprocess RBC volume (mL) RBC depletion (%) MNC recovery (%) Number
COBE2991/Ficoll
Optia/Apheresis
Sepax/Ficoll*
Prodigy/Ficoll
483 6 10 27.5 6 1 133 6 6 18.6 6 2.5 1.2 6 0.5 129 6 14 3.5 6 0.6 97 6 0 98 6 3 3
482 6 10 27.5 6 1 132 6 6 18.6 6 2.5 1.2 6 0.5 53 6 8 3.8 6 0.3 97 6 0 90 6 2 3
100 28.6 28 17.3 1.1 45 0.2 >99 74 2
102 6 3 27.5 6 1 31 6 0.2 18.6 6 2.5 1.2 6 0.5 94 6 14 0.2 6 0.1 >99 65 6 7 3
* One process failure due to defective consumable (tubing set). For the Sepax (n 5 2), mean is shown. For all other data sets (n 5 3), mean6 SEM is shown.
were run in the automatic mode. Hct was tracked using the colorgram from the COBE Spectra device, to maintain an apparent Hct in the collection line of 5%; adjustment of the collection preference to 40 was generally required to warrant that. Once established, the interphase was stable and very few changes in collection preference had to be undertaken. The entire product volume was processed eight or nine times with a typical flow velocity of 120 mL/min and the centrifuge at full speed, which yields a packing factor of approximately 10. The Sepax II with the NeatCell program and corresponding tubing set (CS900.2, Biosafe SA) operated fully automatically as directed by the manufacturer; the device is well established in the cell therapy community.14,15 The CliniMACS Prodigy (Miltenyi Biotec) is a novel multifunctional cell processing automat fitted with a CliniMACS magnet, a centrifuge, a peltier module for heating and cooling, a roller pump, and a number of valves for processing, as well as cameras for automatic surveillance of the centrifuge chamber.25 For the density gradient separation RBC depletion process, of these only pump, centrifuge, and centrifuge surveillance were required. The density gradient separation software was booted and the device was fitted with the RBC depletion consumable as directed by the software. The software subsequently fully automatically performed Ficoll-based RBC depletions, in which it was supported by the centrifuge camera that tracked the MNC layer and adjusted the pumps and valves accordingly. Of note, the density gradient separation package, consisting of the software and the consumable with the centrifuge chamber and tubings, was a prerelease version provided by the manufacturer for evaluation; its public release is impending. Beyond device set-up, density gradient separation on the Prodigy proceeded completely automatically without operator involvement.
Statistical analysis RBC volume was calculated from Hct and product volume. RBC depletion was calculated as the quotient from pre- and postprocess RBC volume. WBC, MNC, or CD341
cell recovery was similarly calculated as pre- and postprocessing total cell numbers. Purity describes the ratio of MNCs to WBCs. Unless otherwise indicated, mean 6 SEM are reported. This being a feasibility analysis, no p values were calculated. Engraftment data for day with absolute neutrophil count of more than 500 3 106/L (neutrophil engraftment) or day with PLT count of more than 20 3 109/L without transfusion (PLT engraftment) were contributed by collaborating clinicians.
RESULTS Feasibility Of the three BC cross-validation runs, all runs were successful except for one Sepax process, where the target population was lost completely due to a defective consumable (tubing set). An oversized plunger had caused a pressure alarm and aborted the process. Moreover, 14, five, 16, and 18 successive BM processings were performed with the COBE2991, the Spectra Optia BMP, the Sepax II, and manual Ficoll, respectively; they proceeded uneventfully and in principle successfully. Of note, all BMs were clinical, not validation products and hence not split and performed as cross-validations. A total of 53 individual marrow samples were processed with the four methods. For large-scale BM processes for BMT, before April 2014, all BM processes were performed with the COBE2991 and with the Spectra Optia BMP thereafter. The Sepax and manual BM processing are part of routine manufacturing of advanced-therapy medicinal products and are used concurrently for different processes, mesenchymal stroma cell generation, and cardiovascular regeneration, respectively.
Quality of processed BC pools Properties of BC pools are given in Table 1. Very robust mean RBC depletions of 97, 97, 99, and 99% were achieved with the COBE2991 (C), Optia (O), Sepax II (S), and Prodigy (P), respectively. The median (range) MNC recoveries were 98% (92%-102%, C), 90% (88%-92%, O), 74% (68%-79%, S), and 65% (58%-69%, P). Starting from a Volume 55, June 2015 TRANSFUSION 1277
SORG ET AL.
from seven recipients of C-processed grafts and all five recipients of O-processed grafts. In these, the mean perkilogram dose of CD341 cells was 2.7 3 106 6 0.6 3 106(C) or 4.0 3 106 6 0.6 3 106 (O). There was no indication of a dose effect; a mean 24 6 3 or 19 6 1 days to neutrophil engraftment and 32 6 8 or 30 6 10 days to PLT engraftment elapsed with C- or O-processed grafts, respectively (Fig. 3).
Time consumption, labor, and costs Time consumption, labor intensity, and costs of the four technologies were also analyzed. Scales displaying total duration and use of labor (two operators during hands-on operation, one during process observation, and none during automatic process steps) are shown in Fig. 4 and allow calculation of cumulative man-hours. With respect to cost, the price of consumables for the COBE2991 and Spectra Optia is approximately half of that for the Sepax II and Prodigy tubing sets. Prorated cost of the device will depend on other uses of the technology besides RBC depletion of BM. Fig. 1. Cross-validation of RBC purging of BC. Large BC pools were generated and unevenly split for concurrent processing by the COBE2991, Spectra Optia BMP, Sepax II NeatCell, or CliniMACS Prodigy density gradient separation. RBC depletion, MNC recovery, and MNC frequency among WBCs (all in %); individual values are shown.
mean MNC frequency of 62% in the BC pools, mean product purities, that is, the median frequencies of MNCs among total WBCs, were 90% (C), 80% (O), 96% (S), and 95% (P; Fig. 1). The higher frequency of non-MNCs with the only Ficoll-free device, the Spectra Optia, is probably less surprising than the considerable (50%) depletion of polymorphonuclear cells that is achieved with the apheresis method. No relevant PLT depletion was observed with the COBE2991 and the Optia; 82 and 66% of PLTs were removed by the Sepax and the Prodigy. Additional quality data for BC processings are given in Table 1. Fifty-three clinical BM samples, sized between 23 and 1511 mL, were processed in total, using the COBE2991, the Spectra Optia, the Sepax II, and manual Ficoll (M). Initial volumes depended on the intended clinical application and technology was selected to accommodate the volume. The technologies removed, on average, 92% (C), 98% (O), or more than 99% (S and M) of RBCs, meanwhile recovering 92% (C), 90% (O), 67% (S), or 70% (M) of CD341 cells (Fig. 2). Full performance data for BM processings are shown in Table 2. Subsequent to RBC depletion, C- and O-processed BM samples were released for transplantation; at 95 6 0.6% (C) or 96 6 0.3% (O), viability of CD341 cells after BM processing with both technologies was excellent. Engraftment reports were received 1278 TRANSFUSION Volume 55, June 2015
DISCUSSION As we are showing, all five technologies generate products that are largely RBC and to varying degree PMN depleted. The almost complete RBC depletion with the Sepax and the Prodigy comes at the price of losing approximately one-third of MNCs (BC) or CD341 cells (BM; data for the Sepax only). The CD341 cell recovery nevertheless exceeded by up to 50% data reported for the Sepax in the literature and thus remained quite satisfactory.14,15 CD341 cell or MNC recoveries with the COBE2991, albeit somewhat variable, and the Optia were near-quantitative, that is, markedly more efficient. The RBC reduction achieved with the COBE2991 and the Optia is less complete than with the Sepax or the Prodigy, but clinically entirely sufficient. Functionality of both products as stem cell grafts is demonstrated by favorable engraftment data. Outcomes for the COBE2991 and the Spectra Optia were in a similar range as those reported for Optia’s predecessor technology, the COBE Spectra.7,27,28 The suitability of BC as a surrogate for BM process validations requires some comment. Given the invasiveness of BM collection, it is clear that the amount of BM available to perform formal process validations is very limited, and in no way can GMP-required performance qualifications of new cell processing technologies in all cell processing laboratories be completed with BM. As is apparent from Tables 1 and 2, Hct (approx. 30%) and WBC (approx. 30 3 109/L) are highly similar in BC and BM. The composition of WBCs in PB and BM, however, obviously differs. MNCs are relatively less frequent in BM and many are not lymphocytes but immature myeloid
NEW RBC DEPLETION TECHNOLOGIES
Fig. 2. RBC depletion from BM. BM depletions were performed with the COBE2991, the Spectra Optia BMP, the Sepax II NeatCell, or manual Ficoll. RBC removal, MNC recovery (Sepax and manual only), and CD341 cell recovery are shown. (䉫) Individual values; bar 5 mean. The respective technology is printed in the graph.
TABLE 2. Performance data of different RBC depletion technologies with BC aspirate* RBC depletion technology Before processing (volume, mL) Preprocess Hct (%) After processing (volume, mL) Postprocess RBC volume (mL) RBC depletion (%) MNC recovery (%) CD341 recovery (%) Number
COBE2991/Ficoll
Optia/Apheresis
Sepax/Ficoll
Manual Ficoll
1268 6 66 28.5 6 1 174 6 19 28.5 6 9.3 91.7 6 2.8 ND 92.3 6 2.1 14
1349 6 42 31.4 6 0.6 195 6 3.6 10.2 6 1.7 97.5 6 0.5 ND 89.5 6 3.3 5
113 6 2.4 33.0 6 1.7 10 6 0 0.09 6 0.01 >99 38 6 6 67 6 12 16
25 6 1 30.2 6 1 11 6 0 0.02 6 0.01 >99 36 6 7 70 6 18 18
* Data are shown as mean 6 SEM. ND 5 not determined.
cells with other physical properties. As our data indicate, RBC depletion in BCs and BMs is equally efficient, and in this respect BC is a suitable replacement for BM. In terms of MNC recovery from BC and BM, differential sedimentation properties become apparent, reflected by very much poorer MNC recoveries in BM than in BC. However, the relevant target cell in BM, the CD341 cell, is recovered by the technologies tested here with nearly the same efficiency from BM as are the MNCs from BC. In other words, technologies associated with good MNC recovery from BM recovered most CD341 cells and vice versa. It is in
our view therefore permissible to perform performance qualifications of the reported technologies with BCs and, positive outcomes provided, then to perform in-process validations with clinical BM aspirates. Whether these observations can be extended to other technologies will need to be individually assessed. Alternatives to RBC depletion, such as in vivo isoagglutinin depletion, have in the past years been developed,29 in part because of concern about CD341 cell loss during BM processing, stem cell dose being one of the strongest predictors of outcome. As we are showing, with Volume 55, June 2015 TRANSFUSION 1279
SORG ET AL.
the COBE2991 and the Optia the mean CD341 cell loss is only 10%, and thus clinically insignificant, but possibly relevant for the other devices. At the current rate of allogeneic BMTs, RBC depletion of marrow will be a lowfrequency process even under the best of circumstances. For reasons of cost-effectiveness, it therefore should probably best be established on a platform not exclusively designated for this process. Successful transplantations with ficolled BM cells have been performed for more than 30 years.13 Since its solubility in water is unlimited, Ficoll is removed quantitatively during the wash steps and the final product is essentially free thereof, but some may remain stuck to cell surfaces. In agreement with clinical observations, the very high viability of BM products processed with Ficoll provides no evidence of adverse effects on cell quality. The cost of Ficoll is modest and irrelevant relative to the total cost for device (depreciation and maintenance), consumable, and labor. Therefore, whether or not the RBC depletion process uses Ficoll or not should likely not guide selection of the technology. The paucity of BMTs in need of RBC depletion and the difficulty of obtaining large-volume volunteer BMs for
Fig. 3. Engraftment data for RBC-depleted BM aspirates. Engraftment for neutrophils (500 3 106/L; 䉫, 䉬) and PLTs (20 3 109/L; 䊊, •) for the COBE2991 (䉫, 䊊) or the Spectra Optia BMP (䉬, •) is plotted by CD341 cell dose (X-axis) over days posttransplant (Y-axis). Both products provided timely engraftment. A dose effect was not apparent.
experimentation does not allow for thorough optimization and validation of BM RBC depletion. It is clear that differential use of inlet flow velocity, packing factor, collection pump flow velocity, and collection preference, all of which are adjustable on the Optia over a wide range, will have significant effects on product properties. Instead, we operated with default settings for inlet flow, packing factor, and collection flow and only adjusted collection preference, so we would collect a product with a Hct of approximately 5% (residual RBC content of 10 mL). With these settings, the Optia recovered on average 90% of CD341 cells, and RBCs were significantly depleted; that is, product quality was quite satisfactory. While theoretically lighter-colored products can be collected with the Optia,16-18,20 this was not attempted: The effort seemed unnecessary because a satisfactory RBC depletion was already achieved with default settings, while a lighter product color would increase the risk of inferior target cell recovery. Similarly, packing factor was not varied from default settings; a higher one might have resulted in a crisper interphase and lighter product, a lower one might have allowed for PLT reduction. However, as we previously documented, expected consequences of variations of apheresis protocols are not always observed, cautioning against overly courageous changes in apheresis variables.17,20 Our data document feasibility and adequate efficiency of RBC depletion while recovering most target cells with the Optia, but attempts at optimization (for, e.g., concurrent PLT depletion) were not made. The capacity for handling of large BM volumes, but also the short process duration, very satisfactory (functional) product properties, and relative automation of Spectra Optia BMP, suggest its use for the typical BM product the volume of which exceeds 300 mL. Availability of a second technology is desirable for processing of smaller volumes of BM such as would be collected from young pediatric donors or from patients for regenerative medicine purposes.15 Moreover, the Spectra Optia requires a minimal RBC volume of 125 mL and the COBE2991 even higher, so that for RBC depletion of less strongly RBC-contaminated products, which nevertheless require further RBC reduction, for example, poorly collected donor lymphocyte infusion products slated for
Fig. 4. Process duration and operator use. Total process length and mode of operation over time for the different technologies are depicted. The arrow indicates the time point where the operator must press the “start” button. 1280 TRANSFUSION Volume 55, June 2015
NEW RBC DEPLETION TECHNOLOGIES
further processing or cultivation, the Sepax or the Prodigy would be the technology of choice. During the BC cross-validation, we observed one system failure (defective tubing set) with the Sepax, where the cell product was lost. While this is a potentially dramatic occurrence given the uniqueness of the material, over the years we have performed several dozen NeatCell processes without any such unfavorable observations, so that we consider the recent singular instrument failure an anomaly and do not intend to discourage its use, in agreement with published data.14,15 MNC recovery and RBC depletion with the Prodigy, which performed flawlessly, were in a similar order of magnitude as Sepax. Full automation is an attractive feature with respect to process documentation and use of labor. Relevant differences in outcome between the two automatic devices were not observed and preference will depend on machine availability. Possible advantages of density gradient separation with the Prodigy as tested here include the potential to handle larger volumes of up to 299 mL, as opposed to the Sepax (120 mL), and the possibility to integrate the Ficoll module into a larger manufacturing process such as subsequent immunomagnetic manipulation, incubation, and so forth. The novel, previously not described RBC depletion technologies, the Spectra Optia BMP and the CliniMACS Prodigy density gradient separation module, are robust and effective. Near-quantitative RBC depletion and CD341 or MNC recovery with the Spectra Optia recommend this technology for large-volume BM cell processing. The Prodigy, like the Sepax, is suitable for processing of smaller volumes, despite somewhat lower target cell recoveries. Unlike the Optia, the latter two will also be suitable for RBC depletion of high-RBC leukapheresis products (e.g., poorly collected donor lymphocyte infusions), a process that is already routinely and successfully performed with the Sepax. ACKNOWLEDGMENT Collaborating clinicians are acknowledged for contributing recipient outcome data.
2. Passweg JR, Baldomero H, Bregni M, et al. Hematopoietic SCT in Europe: data and trends in 2011. Bone Marrow Transplant 2013;48:1161-7. 3. Barrett AJ, Ringd en O, Zhang MJ, et al. Effect of nucleated marrow cell dose on relapse and survival in identical twin bone marrow transplants for leukemia. Blood 2000;95:3323-7. 4. Dominietto A, Lamparelli T, Raiola AM, et al. Transplantrelated mortality and long-term graft function are significantly influenced by cell dose in patients undergoing allogeneic marrow transplantation. Blood 2002;100:3930-4. 5. Sierra J, Storer B, Hansen JA, et al. Transplantation of marrow cells from unrelated donors for treatment of high-risk acute leukemia: the effect of leukemic burden, donor HLAmatching, and marrow cell dose. Blood 1997;89:4226-35. 6. Blacklock HA, Gilmore MJ, Prentice HG, et al. ABO-incompatible bone-marrow transplantation: removal of red blood cells from donor marrow avoiding recipient antibody depletion. Lancet 1982;2:1061-4. 7. Korıstek Z, Mayer J. Bone marrow processing for transplantation using the COBE spectra cell separator. J Hematother Stem Cell Res 1999;8:443-8. €ck E, Fritsch G. Bone marrow processing with 8. Witt V, Beiglbo the AMICUS separator system. J Clin Apher 2011;26:195-9. 9. Menichella G, Pierelli L, Dragani A, et al. A preliminary survey of Italian experience on bone marrow harvesting, processing and manipulation. The Italian Cooperative Study Group on Cell Manipulation in Hematology. Haematologica 1990;75:39-42. 10. Boyum A. Isolation of mononuclear cells and granulocytes from human blood. Isolation of mononuclear cells by one centrifugation, and of granulocytes by combining centrifugation and sedimentation at 1 g. Scand J Clin Lab Invest Suppl 1968;97:77-89. 11. Daniel-Johnson J, Schwartz J. How do I approach ABOincompatible hematopoietic progenitor cell transplantation? Transfusion 2011;51:1143-9. 12. Jin NR, Hill R, Segal G, et al. Preparation of red-blood-celldepleted marrow for ABO-incompatible marrow transplantation by density-gradient separation using the IBM 2991 blood cell processor. Exp Hematol 1987;15:93-8. 13. Preti RA, Ahmed T, Ayello J, et al. Hemopoietic stem cell processing: comparison of progenitor cell recovery using the
CONFLICT OF INTEREST HBB has received honoraria (speaker’s board) and/or research support from Miltenyi Biotech and Terumo BCT (formerly Caridian BCT), manufacturers of technology tested here. None of the other authors have disclosed any conflicts of interest.
REFERENCES 1. Blin N, Traineau R, Houssin S, et al. Impact of donorrecipient major ABO mismatch on allogeneic transplantation outcome according to stem cell source. Biol Blood Marrow Transplant 2010;16:1315-23.
Cobe 2991 cell washer and the Haemonetics V50 apheresis system. Bone Marrow Transplant 1992;9:377-81. 14. Aktas M, Radke TF, Strauer BE, et al. Separation of adult bone marrow mononuclear cells using the automated closed separation system Sepax. Cytotherapy 2008;10:203-11. 15. Gee AP, Richman S, Durett A, et al. Multicenter cell processing for cardiovascular regenerative medicine applications: the Cardiovascular Cell Therapy Research Network (CCTRN) experience. Cytotherapy 2010;12:684-91. 16. Brauninger S, Bialleck H, Thorausch K, et al. Mobilized allogeneic peripheral stem/progenitor cell apheresis with Spectra Optia v.5.0, a novel, automatic interface-controlled apheresis system: results from the first feasibility trial. Vox Sang 2011;101:237-46. Volume 55, June 2015 TRANSFUSION 1281
SORG ET AL.
17. Brauninger S, Bialleck H, Thorausch K, et al. Allogeneic donor peripheral blood “stem cell” apheresis: prospective comparison of two apheresis systems. Transfusion 2012;52:1137-45. 18. Reinhardt P, Brauninger S, Bialleck H, et al. Automatic
24. Thorausch K, Schulz M, Bialleck H, et al. Granulocyte collections: comparison of two apheresis systems. Transfusion 2013;53:3262-8. 25. Apel M, Bruning M, Granzin M, et al. Integrated clinical scale
interface-controlled apheresis collection of stem/progenitor
manufacturing system for cellular products derived by mag-
cells: results from an autologous donor validation trial of a novel stem cell apheresis device. Transfusion 2011;51:1321-
netic cell separation, centrifugation and cell culture. Chemie Ingenieur Technik 2014;85:103-10.
30. 19. Schulz M, Bug G, Bialleck H, et al. Leucodepletion for hyperleucocytosis—first report on a novel technology featuring electronic interphase management. Vox Sang 2013;105:47-53. 20. Schulz M, Bialleck H, Thorausch K, et al. Unstimulated leukapheresis in patients and donors: comparison of two apheresis systems. Transfusion 2014;54:1622-9. €rensen J, Jarisch A, Smorta C, et al. Pediatric apheresis 21. So with a novel apheresis device with electronic interface control. Transfusion 2013;53:761-5. 22. Steininger PA, Smith R, Geier C, et al. Leukapheresis in noncytokine-stimulated donors with a new apheresis system: first-time collection results and evaluation of subsequent cryopreservation. Transfusion 2013;53:747-56. 23. Steininger PA, Strasser EF, Weiss D, et al. First comparative
26. Dauber K, Becker D, Odendahl M, et al. Enumeration of viable CD34(1) cells by flow cytometry in blood, bone marrow and cord blood: results of a study of the novel BD stem cell enumeration kit. Cytotherapy 2011;13:449-58. 27. Guttridge MG, Sidders C, Booth-Davey E, et al. Factors affecting volume reduction and red blood cell depletion of bone marrow on the COBE Spectra cell separator before haematopoietic stem cell transplantation. Bone Marrow Transplant 2006;38:175-81. 28. Larghero J, Rea D, Esperou H, et al. ABO-mismatched marrow processing for transplantation: results of 114 procedures and analysis of immediate adverse events and hematopoietic recovery. Transfusion 2006;46:398-402. 29. Scholl S, Klink A, Mugge LO, et al. Safety and impact of donor-type red blood cell transfusion before allogeneic
evaluation of a new leukapheresis technology in non-
peripheral blood progenitor cell transplantation with major
cytokine-stimulated donors. Vox Sang 2014;106:248-55.
ABO mismatch. Transfusion 2005;45:1676-83.
1282 TRANSFUSION Volume 55, June 2015
BACKGROUND: Red blood cell (RBC) depletion is a standard technique for preparation of ABO-incompatible bone marrow transplants (BMTs). Density centrifugation or apheresis are used successfully at clinical scale. The advent of a bone marrow (BM) processing module for the Spectra Optia (Terumo BCT) provided the initiative to formally compare our standard technology, the COBE2991 (Ficoll, manual, “C”) with the Spectra Optia BMP (apheresis, semiautomatic, “O”), the Sepax II NeatCell (Ficoll, automatic, “S”), the Miltenyi CliniMACS Prodigy density gradient separation system (Ficoll, automatic, “P”), and manual Ficoll (“M”). C and O handle larger product volumes than S, P, and M. STUDY DESIGN AND METHODS: Technologies were assessed for RBC depletion, target cell (mononuclear cells [MNCs] for buffy coats [BCs], CD341 cells for BM) recovery, and cost/labor. BC pools were simultaneously purged with C, O, S, and P; five to 18 BM samples were sequentially processed with C, O, S, and M. RESULTS: Mean RBC removal with C was 97% (BCs) or 92% (BM). From both products, O removed 97%, and P, S, and M removed 99% of RBCs. MNC recovery from BC (98% C, 97% O, 65% P, 74% S) or CD341 cell recovery from BM (92% C, 90% O, 67% S, 70% M) were best with C and O. Polymorphonuclear cells (PMNs) were depleted from BCs by P, S, and C, while O recovered 50% of PMNs. Time savings compared to C or M for all tested technologies are considerable. CONCLUSION: All methods are in principle suitable and can be selected based on sample volume, available technology, and desired product specifications beyond RBC depletion and MNC and/or CD341 cell recovery.
S
ince HLA and A-/B-transferase genes are inherited independently, an ABO major or bidirectional mismatch is observed in roughly 20% of bone marrow transplantations (BMTs),1 that is, approximately 600 times per year in all of Europe with its more than 600 transplant programs2 or once per year per center. A typical dose of bone marrow (BM) cells for BMT is 2.5 3 108 to 3.0 3 108/kg nucleated cells,3-5 equivalent to approximately 12 to 20 mL of BM aspirate/kg. At a mean hematocrit (Hct) of 30% to 35%, the red blood cell (RBC) content of BM aspirate is thus considerable, 4 to 7 mL/kg of the recipient, and not unconditionally tolerable for ABO major–mismatched recipients. One long-established method to avoid severe and potentially lethal hemolysis ABBREVIATIONS: BC 5 buffy coat; BM 5 bone marrow. From the 1German Red Cross Blood Service Baden€rttemberg-Hessen, Institute Frankfurt, and the 2Institute for Wu Transfusion Medicine and Immunohematology, Goethe University, Frankfurt, Germany; and the 3Department of Medicine/Hematology, University of Washington, Seattle, Washington. € nig, Goethe University Address reprint requests to: Halvard Bo Medical Center, Institute for Transfusion Medicine and Immunohematology, and German Red Cross Blood Service Baden€rttemberg-Hesse, Sandhofstrasse 1, 60528 Frankfurt, Germany; Wu e-mail: [email protected] [email protected]. *Contributed equally. HB and ES are members of the LOEWE Cell and Gene Therapy Frankfurt faculty funded by Hessian Ministry of Higher Education, Research and the Arts ref.no.: III L 4-518/ 17.004 (2010). RBC-depletion tubing sets for Prodigy were gifted by Miltenyi Biotech. No outside funding was used to support the studies. Received for publication September 24, 2014; revision received November 24, 2014; and accepted November 28, 2014. doi:10.1111/trf.13001 C 2015 AABB V
TRANSFUSION 2015;55;1275–1282 Volume 55, June 2015 TRANSFUSION 1275
SORG ET AL.
during BM infusion is RBC depletion from the graft.6 RBC depletion is also used for volume reduction of BM aspirate before cryoconservation. Since RBCs, white blood cells (WBCs), platelets (PLTs), and plasma differ in size and density, they can be efficiently separated by density gradient centrifugation or apheresis, and both have been applied successfully.6-9 The critical variables for the success of RBC depletion of BM are recovery of CD341 “stem cells,” which are contained within the mononuclear cell (MNC) population and, when performed in the context of ABO major– incompatible BMT, near-qualitative removal of RBCs to values below 0.5 mL/kg body weight of the recipient. Similar processing of whole blood, buffy coats (BCs), and occasionally, apheresis products generally also aims for best-possible RBC depletion while maximally recovering MNCs. Density gradient centrifugation with “Ficoll,” a synthetic low-viscosity, high-molecular-weight hydrophilic sucrose, and epichlorohydrine polymer and sodium diatrizoate, has been extensively used for RBC (and PMN) depletion.10 Alternatively, apheresis technology has been applied to deplete RBCs from BM. The modest differences in separation physics between density gradient centrifugation and apheresis allow some prediction of RBC depletion outcomes. In terms of available technology, the COBE29916,11-13 (Terumo BCT, Lakewood, CO) and the Sepax II NeatCell14,15 (Biosafe SA, Eysins, Switzerland) have been on the market for many decades and several years, respectively. Both are based on Ficoll density centrifugation, and significant evidence of their suitability has been provided. The Spectra Optia BMP is a new application for the relatively recent Spectra Optia apheresis device that was previously introduced.16-24 The package consists of the standard Spectra Optia MNC filler, the standard Spectra Optia IDL tubing set, and a large BM bag for recirculating of the (increasingly MNC-deplete) BM suspension, as well as a software that provides fully automatic, photo detector– guided apheresis of the BM, physically highly analogous to MNC apheresis in stem cell donors. The Spectra Optia BMP is licensed in the European Medicines Agency region, but not in the United States. The CliniMACS Prodigy (Miltenyi Biotec, Bergisch Gladbach, Germany) device is a multifunctional automatic cell processing device consisting of, among other features, a camera-fitted centrifugation chamber, where the device can locate, track, and selectively aspirate the MNC layer; a roller pump; valves; and a single-use sterile all-in-one consumable.25 The Ficoll package for Prodigy has not yet been publically released; a near-series beta version of the software and prototypic plasticware provided by the manufacturer were used for the validation. For neither of the two latter methods have performance data been reported. In a side-byside validation, large BC pools from 20 donors each were 1276 TRANSFUSION Volume 55, June 2015
split and subjected to concurrent RBC depletion with all four technologies. Moreover, 53 clinical BM aspirates were RBC depleted with the COBE2991, the Spectra Optia BMP, the Sepax II NeatCell, or manual Ficoll. Results indicating feasibility and clinically adequate efficiency, albeit with some quantitative variation, are presented.
MATERIALS AND METHODS BC pools were generated from residual BCs from 20 healthy, ABO-identical whole blood donors with donor and internal review board approval (permit number 329/ 19). Two large (approx. 500 mL) and two small (100 mL) aliquots were concurrently separated with four different technologies. BM samples were clinical BM aspirates for hematopoietic reconstitution (COBE2991, n 5 14; Spectra Optia, n 5 5), mesenchymal stroma cell isolation (Sepax, n 5 16), or cardiovascular regeneration (manual Ficoll, n 5 18); routine performance data were collected after pseudonymization. Product quality including Hct was assessed by automatic hemocytometry (Sysmex XT-1800i hematology analyzer, Sysmex, Norderstedt, Germany); in the processed marrow samples, CD341 cells were enumerated using the single-platform stem cell enumeration kit and flow cytometer (FACSCalibur, Becton-Dickinson, Heidelberg, Germany), as described.26 Volume was assessed by weight with correction for Hct. Reagents required for the RBC depletions include density medium (Ficoll, Lonza, Verviers, Belgium), as well as NaCl 0.9% and human serum albumin (Baxter Healthcare, Unterschleißheim, Germany).
Separation devices and protocols One manual (COBE2991), one semiautomatic (Spectra Optia) and two fully automatic, walk-away (Sepax II and CliniMACS Prodigy) RBC depletion protocols were used, in addition to fully manual semiopen Ficoll processing in 50-mL conical tubes. The COBE2991 device was fitted with the corresponding consumable (Terumo BCT); the manual process was performed according to standard operating procedures based on the manufacturer’s instructions. Briefly, separation of WBCs and RBCs was achieved by layering Ficoll under the BC in the centrifugation bag (donut bag) and extraction of the different phases by inflating the cushion underneath and manual clamping or unclamping of the waste, nontarget, and target bags based on visual detection of the different fractions in the tubing set. The Spectra Optia, previously described with respect to its performance as patient or donor apheresis device,16-24 was fitted with the standard filler (as for MNC apheresis), the IDL separation kit, and the BMP accessory kit. The set was installed and primed as directed by the manufacturer. An initial collection preference16 of 50 was selected, as manufacturer recommended; RBC depletions
NEW RBC DEPLETION TECHNOLOGIES
TABLE 1. Performance data of different RBC depletion technologies with BC pools RBC depletion technology Before processing (volume, mL) Preprocess Hct (%) Preprocess RBC volume (mL) Preprocess MNC count (3109/L) PLT count (3109/L) After processing (volume, mL) Postprocess RBC volume (mL) RBC depletion (%) MNC recovery (%) Number
COBE2991/Ficoll
Optia/Apheresis
Sepax/Ficoll*
Prodigy/Ficoll
483 6 10 27.5 6 1 133 6 6 18.6 6 2.5 1.2 6 0.5 129 6 14 3.5 6 0.6 97 6 0 98 6 3 3
482 6 10 27.5 6 1 132 6 6 18.6 6 2.5 1.2 6 0.5 53 6 8 3.8 6 0.3 97 6 0 90 6 2 3
100 28.6 28 17.3 1.1 45 0.2 >99 74 2
102 6 3 27.5 6 1 31 6 0.2 18.6 6 2.5 1.2 6 0.5 94 6 14 0.2 6 0.1 >99 65 6 7 3
* One process failure due to defective consumable (tubing set). For the Sepax (n 5 2), mean is shown. For all other data sets (n 5 3), mean6 SEM is shown.
were run in the automatic mode. Hct was tracked using the colorgram from the COBE Spectra device, to maintain an apparent Hct in the collection line of 5%; adjustment of the collection preference to 40 was generally required to warrant that. Once established, the interphase was stable and very few changes in collection preference had to be undertaken. The entire product volume was processed eight or nine times with a typical flow velocity of 120 mL/min and the centrifuge at full speed, which yields a packing factor of approximately 10. The Sepax II with the NeatCell program and corresponding tubing set (CS900.2, Biosafe SA) operated fully automatically as directed by the manufacturer; the device is well established in the cell therapy community.14,15 The CliniMACS Prodigy (Miltenyi Biotec) is a novel multifunctional cell processing automat fitted with a CliniMACS magnet, a centrifuge, a peltier module for heating and cooling, a roller pump, and a number of valves for processing, as well as cameras for automatic surveillance of the centrifuge chamber.25 For the density gradient separation RBC depletion process, of these only pump, centrifuge, and centrifuge surveillance were required. The density gradient separation software was booted and the device was fitted with the RBC depletion consumable as directed by the software. The software subsequently fully automatically performed Ficoll-based RBC depletions, in which it was supported by the centrifuge camera that tracked the MNC layer and adjusted the pumps and valves accordingly. Of note, the density gradient separation package, consisting of the software and the consumable with the centrifuge chamber and tubings, was a prerelease version provided by the manufacturer for evaluation; its public release is impending. Beyond device set-up, density gradient separation on the Prodigy proceeded completely automatically without operator involvement.
Statistical analysis RBC volume was calculated from Hct and product volume. RBC depletion was calculated as the quotient from pre- and postprocess RBC volume. WBC, MNC, or CD341
cell recovery was similarly calculated as pre- and postprocessing total cell numbers. Purity describes the ratio of MNCs to WBCs. Unless otherwise indicated, mean 6 SEM are reported. This being a feasibility analysis, no p values were calculated. Engraftment data for day with absolute neutrophil count of more than 500 3 106/L (neutrophil engraftment) or day with PLT count of more than 20 3 109/L without transfusion (PLT engraftment) were contributed by collaborating clinicians.
RESULTS Feasibility Of the three BC cross-validation runs, all runs were successful except for one Sepax process, where the target population was lost completely due to a defective consumable (tubing set). An oversized plunger had caused a pressure alarm and aborted the process. Moreover, 14, five, 16, and 18 successive BM processings were performed with the COBE2991, the Spectra Optia BMP, the Sepax II, and manual Ficoll, respectively; they proceeded uneventfully and in principle successfully. Of note, all BMs were clinical, not validation products and hence not split and performed as cross-validations. A total of 53 individual marrow samples were processed with the four methods. For large-scale BM processes for BMT, before April 2014, all BM processes were performed with the COBE2991 and with the Spectra Optia BMP thereafter. The Sepax and manual BM processing are part of routine manufacturing of advanced-therapy medicinal products and are used concurrently for different processes, mesenchymal stroma cell generation, and cardiovascular regeneration, respectively.
Quality of processed BC pools Properties of BC pools are given in Table 1. Very robust mean RBC depletions of 97, 97, 99, and 99% were achieved with the COBE2991 (C), Optia (O), Sepax II (S), and Prodigy (P), respectively. The median (range) MNC recoveries were 98% (92%-102%, C), 90% (88%-92%, O), 74% (68%-79%, S), and 65% (58%-69%, P). Starting from a Volume 55, June 2015 TRANSFUSION 1277
SORG ET AL.
from seven recipients of C-processed grafts and all five recipients of O-processed grafts. In these, the mean perkilogram dose of CD341 cells was 2.7 3 106 6 0.6 3 106(C) or 4.0 3 106 6 0.6 3 106 (O). There was no indication of a dose effect; a mean 24 6 3 or 19 6 1 days to neutrophil engraftment and 32 6 8 or 30 6 10 days to PLT engraftment elapsed with C- or O-processed grafts, respectively (Fig. 3).
Time consumption, labor, and costs Time consumption, labor intensity, and costs of the four technologies were also analyzed. Scales displaying total duration and use of labor (two operators during hands-on operation, one during process observation, and none during automatic process steps) are shown in Fig. 4 and allow calculation of cumulative man-hours. With respect to cost, the price of consumables for the COBE2991 and Spectra Optia is approximately half of that for the Sepax II and Prodigy tubing sets. Prorated cost of the device will depend on other uses of the technology besides RBC depletion of BM. Fig. 1. Cross-validation of RBC purging of BC. Large BC pools were generated and unevenly split for concurrent processing by the COBE2991, Spectra Optia BMP, Sepax II NeatCell, or CliniMACS Prodigy density gradient separation. RBC depletion, MNC recovery, and MNC frequency among WBCs (all in %); individual values are shown.
mean MNC frequency of 62% in the BC pools, mean product purities, that is, the median frequencies of MNCs among total WBCs, were 90% (C), 80% (O), 96% (S), and 95% (P; Fig. 1). The higher frequency of non-MNCs with the only Ficoll-free device, the Spectra Optia, is probably less surprising than the considerable (50%) depletion of polymorphonuclear cells that is achieved with the apheresis method. No relevant PLT depletion was observed with the COBE2991 and the Optia; 82 and 66% of PLTs were removed by the Sepax and the Prodigy. Additional quality data for BC processings are given in Table 1. Fifty-three clinical BM samples, sized between 23 and 1511 mL, were processed in total, using the COBE2991, the Spectra Optia, the Sepax II, and manual Ficoll (M). Initial volumes depended on the intended clinical application and technology was selected to accommodate the volume. The technologies removed, on average, 92% (C), 98% (O), or more than 99% (S and M) of RBCs, meanwhile recovering 92% (C), 90% (O), 67% (S), or 70% (M) of CD341 cells (Fig. 2). Full performance data for BM processings are shown in Table 2. Subsequent to RBC depletion, C- and O-processed BM samples were released for transplantation; at 95 6 0.6% (C) or 96 6 0.3% (O), viability of CD341 cells after BM processing with both technologies was excellent. Engraftment reports were received 1278 TRANSFUSION Volume 55, June 2015
DISCUSSION As we are showing, all five technologies generate products that are largely RBC and to varying degree PMN depleted. The almost complete RBC depletion with the Sepax and the Prodigy comes at the price of losing approximately one-third of MNCs (BC) or CD341 cells (BM; data for the Sepax only). The CD341 cell recovery nevertheless exceeded by up to 50% data reported for the Sepax in the literature and thus remained quite satisfactory.14,15 CD341 cell or MNC recoveries with the COBE2991, albeit somewhat variable, and the Optia were near-quantitative, that is, markedly more efficient. The RBC reduction achieved with the COBE2991 and the Optia is less complete than with the Sepax or the Prodigy, but clinically entirely sufficient. Functionality of both products as stem cell grafts is demonstrated by favorable engraftment data. Outcomes for the COBE2991 and the Spectra Optia were in a similar range as those reported for Optia’s predecessor technology, the COBE Spectra.7,27,28 The suitability of BC as a surrogate for BM process validations requires some comment. Given the invasiveness of BM collection, it is clear that the amount of BM available to perform formal process validations is very limited, and in no way can GMP-required performance qualifications of new cell processing technologies in all cell processing laboratories be completed with BM. As is apparent from Tables 1 and 2, Hct (approx. 30%) and WBC (approx. 30 3 109/L) are highly similar in BC and BM. The composition of WBCs in PB and BM, however, obviously differs. MNCs are relatively less frequent in BM and many are not lymphocytes but immature myeloid
NEW RBC DEPLETION TECHNOLOGIES
Fig. 2. RBC depletion from BM. BM depletions were performed with the COBE2991, the Spectra Optia BMP, the Sepax II NeatCell, or manual Ficoll. RBC removal, MNC recovery (Sepax and manual only), and CD341 cell recovery are shown. (䉫) Individual values; bar 5 mean. The respective technology is printed in the graph.
TABLE 2. Performance data of different RBC depletion technologies with BC aspirate* RBC depletion technology Before processing (volume, mL) Preprocess Hct (%) After processing (volume, mL) Postprocess RBC volume (mL) RBC depletion (%) MNC recovery (%) CD341 recovery (%) Number
COBE2991/Ficoll
Optia/Apheresis
Sepax/Ficoll
Manual Ficoll
1268 6 66 28.5 6 1 174 6 19 28.5 6 9.3 91.7 6 2.8 ND 92.3 6 2.1 14
1349 6 42 31.4 6 0.6 195 6 3.6 10.2 6 1.7 97.5 6 0.5 ND 89.5 6 3.3 5
113 6 2.4 33.0 6 1.7 10 6 0 0.09 6 0.01 >99 38 6 6 67 6 12 16
25 6 1 30.2 6 1 11 6 0 0.02 6 0.01 >99 36 6 7 70 6 18 18
* Data are shown as mean 6 SEM. ND 5 not determined.
cells with other physical properties. As our data indicate, RBC depletion in BCs and BMs is equally efficient, and in this respect BC is a suitable replacement for BM. In terms of MNC recovery from BC and BM, differential sedimentation properties become apparent, reflected by very much poorer MNC recoveries in BM than in BC. However, the relevant target cell in BM, the CD341 cell, is recovered by the technologies tested here with nearly the same efficiency from BM as are the MNCs from BC. In other words, technologies associated with good MNC recovery from BM recovered most CD341 cells and vice versa. It is in
our view therefore permissible to perform performance qualifications of the reported technologies with BCs and, positive outcomes provided, then to perform in-process validations with clinical BM aspirates. Whether these observations can be extended to other technologies will need to be individually assessed. Alternatives to RBC depletion, such as in vivo isoagglutinin depletion, have in the past years been developed,29 in part because of concern about CD341 cell loss during BM processing, stem cell dose being one of the strongest predictors of outcome. As we are showing, with Volume 55, June 2015 TRANSFUSION 1279
SORG ET AL.
the COBE2991 and the Optia the mean CD341 cell loss is only 10%, and thus clinically insignificant, but possibly relevant for the other devices. At the current rate of allogeneic BMTs, RBC depletion of marrow will be a lowfrequency process even under the best of circumstances. For reasons of cost-effectiveness, it therefore should probably best be established on a platform not exclusively designated for this process. Successful transplantations with ficolled BM cells have been performed for more than 30 years.13 Since its solubility in water is unlimited, Ficoll is removed quantitatively during the wash steps and the final product is essentially free thereof, but some may remain stuck to cell surfaces. In agreement with clinical observations, the very high viability of BM products processed with Ficoll provides no evidence of adverse effects on cell quality. The cost of Ficoll is modest and irrelevant relative to the total cost for device (depreciation and maintenance), consumable, and labor. Therefore, whether or not the RBC depletion process uses Ficoll or not should likely not guide selection of the technology. The paucity of BMTs in need of RBC depletion and the difficulty of obtaining large-volume volunteer BMs for
Fig. 3. Engraftment data for RBC-depleted BM aspirates. Engraftment for neutrophils (500 3 106/L; 䉫, 䉬) and PLTs (20 3 109/L; 䊊, •) for the COBE2991 (䉫, 䊊) or the Spectra Optia BMP (䉬, •) is plotted by CD341 cell dose (X-axis) over days posttransplant (Y-axis). Both products provided timely engraftment. A dose effect was not apparent.
experimentation does not allow for thorough optimization and validation of BM RBC depletion. It is clear that differential use of inlet flow velocity, packing factor, collection pump flow velocity, and collection preference, all of which are adjustable on the Optia over a wide range, will have significant effects on product properties. Instead, we operated with default settings for inlet flow, packing factor, and collection flow and only adjusted collection preference, so we would collect a product with a Hct of approximately 5% (residual RBC content of 10 mL). With these settings, the Optia recovered on average 90% of CD341 cells, and RBCs were significantly depleted; that is, product quality was quite satisfactory. While theoretically lighter-colored products can be collected with the Optia,16-18,20 this was not attempted: The effort seemed unnecessary because a satisfactory RBC depletion was already achieved with default settings, while a lighter product color would increase the risk of inferior target cell recovery. Similarly, packing factor was not varied from default settings; a higher one might have resulted in a crisper interphase and lighter product, a lower one might have allowed for PLT reduction. However, as we previously documented, expected consequences of variations of apheresis protocols are not always observed, cautioning against overly courageous changes in apheresis variables.17,20 Our data document feasibility and adequate efficiency of RBC depletion while recovering most target cells with the Optia, but attempts at optimization (for, e.g., concurrent PLT depletion) were not made. The capacity for handling of large BM volumes, but also the short process duration, very satisfactory (functional) product properties, and relative automation of Spectra Optia BMP, suggest its use for the typical BM product the volume of which exceeds 300 mL. Availability of a second technology is desirable for processing of smaller volumes of BM such as would be collected from young pediatric donors or from patients for regenerative medicine purposes.15 Moreover, the Spectra Optia requires a minimal RBC volume of 125 mL and the COBE2991 even higher, so that for RBC depletion of less strongly RBC-contaminated products, which nevertheless require further RBC reduction, for example, poorly collected donor lymphocyte infusion products slated for
Fig. 4. Process duration and operator use. Total process length and mode of operation over time for the different technologies are depicted. The arrow indicates the time point where the operator must press the “start” button. 1280 TRANSFUSION Volume 55, June 2015
NEW RBC DEPLETION TECHNOLOGIES
further processing or cultivation, the Sepax or the Prodigy would be the technology of choice. During the BC cross-validation, we observed one system failure (defective tubing set) with the Sepax, where the cell product was lost. While this is a potentially dramatic occurrence given the uniqueness of the material, over the years we have performed several dozen NeatCell processes without any such unfavorable observations, so that we consider the recent singular instrument failure an anomaly and do not intend to discourage its use, in agreement with published data.14,15 MNC recovery and RBC depletion with the Prodigy, which performed flawlessly, were in a similar order of magnitude as Sepax. Full automation is an attractive feature with respect to process documentation and use of labor. Relevant differences in outcome between the two automatic devices were not observed and preference will depend on machine availability. Possible advantages of density gradient separation with the Prodigy as tested here include the potential to handle larger volumes of up to 299 mL, as opposed to the Sepax (120 mL), and the possibility to integrate the Ficoll module into a larger manufacturing process such as subsequent immunomagnetic manipulation, incubation, and so forth. The novel, previously not described RBC depletion technologies, the Spectra Optia BMP and the CliniMACS Prodigy density gradient separation module, are robust and effective. Near-quantitative RBC depletion and CD341 or MNC recovery with the Spectra Optia recommend this technology for large-volume BM cell processing. The Prodigy, like the Sepax, is suitable for processing of smaller volumes, despite somewhat lower target cell recoveries. Unlike the Optia, the latter two will also be suitable for RBC depletion of high-RBC leukapheresis products (e.g., poorly collected donor lymphocyte infusions), a process that is already routinely and successfully performed with the Sepax. ACKNOWLEDGMENT Collaborating clinicians are acknowledged for contributing recipient outcome data.
2. Passweg JR, Baldomero H, Bregni M, et al. Hematopoietic SCT in Europe: data and trends in 2011. Bone Marrow Transplant 2013;48:1161-7. 3. Barrett AJ, Ringd en O, Zhang MJ, et al. Effect of nucleated marrow cell dose on relapse and survival in identical twin bone marrow transplants for leukemia. Blood 2000;95:3323-7. 4. Dominietto A, Lamparelli T, Raiola AM, et al. Transplantrelated mortality and long-term graft function are significantly influenced by cell dose in patients undergoing allogeneic marrow transplantation. Blood 2002;100:3930-4. 5. Sierra J, Storer B, Hansen JA, et al. Transplantation of marrow cells from unrelated donors for treatment of high-risk acute leukemia: the effect of leukemic burden, donor HLAmatching, and marrow cell dose. Blood 1997;89:4226-35. 6. Blacklock HA, Gilmore MJ, Prentice HG, et al. ABO-incompatible bone-marrow transplantation: removal of red blood cells from donor marrow avoiding recipient antibody depletion. Lancet 1982;2:1061-4. 7. Korıstek Z, Mayer J. Bone marrow processing for transplantation using the COBE spectra cell separator. J Hematother Stem Cell Res 1999;8:443-8. €ck E, Fritsch G. Bone marrow processing with 8. Witt V, Beiglbo the AMICUS separator system. J Clin Apher 2011;26:195-9. 9. Menichella G, Pierelli L, Dragani A, et al. A preliminary survey of Italian experience on bone marrow harvesting, processing and manipulation. The Italian Cooperative Study Group on Cell Manipulation in Hematology. Haematologica 1990;75:39-42. 10. Boyum A. Isolation of mononuclear cells and granulocytes from human blood. Isolation of mononuclear cells by one centrifugation, and of granulocytes by combining centrifugation and sedimentation at 1 g. Scand J Clin Lab Invest Suppl 1968;97:77-89. 11. Daniel-Johnson J, Schwartz J. How do I approach ABOincompatible hematopoietic progenitor cell transplantation? Transfusion 2011;51:1143-9. 12. Jin NR, Hill R, Segal G, et al. Preparation of red-blood-celldepleted marrow for ABO-incompatible marrow transplantation by density-gradient separation using the IBM 2991 blood cell processor. Exp Hematol 1987;15:93-8. 13. Preti RA, Ahmed T, Ayello J, et al. Hemopoietic stem cell processing: comparison of progenitor cell recovery using the
CONFLICT OF INTEREST HBB has received honoraria (speaker’s board) and/or research support from Miltenyi Biotech and Terumo BCT (formerly Caridian BCT), manufacturers of technology tested here. None of the other authors have disclosed any conflicts of interest.
REFERENCES 1. Blin N, Traineau R, Houssin S, et al. Impact of donorrecipient major ABO mismatch on allogeneic transplantation outcome according to stem cell source. Biol Blood Marrow Transplant 2010;16:1315-23.
Cobe 2991 cell washer and the Haemonetics V50 apheresis system. Bone Marrow Transplant 1992;9:377-81. 14. Aktas M, Radke TF, Strauer BE, et al. Separation of adult bone marrow mononuclear cells using the automated closed separation system Sepax. Cytotherapy 2008;10:203-11. 15. Gee AP, Richman S, Durett A, et al. Multicenter cell processing for cardiovascular regenerative medicine applications: the Cardiovascular Cell Therapy Research Network (CCTRN) experience. Cytotherapy 2010;12:684-91. 16. Brauninger S, Bialleck H, Thorausch K, et al. Mobilized allogeneic peripheral stem/progenitor cell apheresis with Spectra Optia v.5.0, a novel, automatic interface-controlled apheresis system: results from the first feasibility trial. Vox Sang 2011;101:237-46. Volume 55, June 2015 TRANSFUSION 1281
SORG ET AL.
17. Brauninger S, Bialleck H, Thorausch K, et al. Allogeneic donor peripheral blood “stem cell” apheresis: prospective comparison of two apheresis systems. Transfusion 2012;52:1137-45. 18. Reinhardt P, Brauninger S, Bialleck H, et al. Automatic
24. Thorausch K, Schulz M, Bialleck H, et al. Granulocyte collections: comparison of two apheresis systems. Transfusion 2013;53:3262-8. 25. Apel M, Bruning M, Granzin M, et al. Integrated clinical scale
interface-controlled apheresis collection of stem/progenitor
manufacturing system for cellular products derived by mag-
cells: results from an autologous donor validation trial of a novel stem cell apheresis device. Transfusion 2011;51:1321-
netic cell separation, centrifugation and cell culture. Chemie Ingenieur Technik 2014;85:103-10.
30. 19. Schulz M, Bug G, Bialleck H, et al. Leucodepletion for hyperleucocytosis—first report on a novel technology featuring electronic interphase management. Vox Sang 2013;105:47-53. 20. Schulz M, Bialleck H, Thorausch K, et al. Unstimulated leukapheresis in patients and donors: comparison of two apheresis systems. Transfusion 2014;54:1622-9. €rensen J, Jarisch A, Smorta C, et al. Pediatric apheresis 21. So with a novel apheresis device with electronic interface control. Transfusion 2013;53:761-5. 22. Steininger PA, Smith R, Geier C, et al. Leukapheresis in noncytokine-stimulated donors with a new apheresis system: first-time collection results and evaluation of subsequent cryopreservation. Transfusion 2013;53:747-56. 23. Steininger PA, Strasser EF, Weiss D, et al. First comparative
26. Dauber K, Becker D, Odendahl M, et al. Enumeration of viable CD34(1) cells by flow cytometry in blood, bone marrow and cord blood: results of a study of the novel BD stem cell enumeration kit. Cytotherapy 2011;13:449-58. 27. Guttridge MG, Sidders C, Booth-Davey E, et al. Factors affecting volume reduction and red blood cell depletion of bone marrow on the COBE Spectra cell separator before haematopoietic stem cell transplantation. Bone Marrow Transplant 2006;38:175-81. 28. Larghero J, Rea D, Esperou H, et al. ABO-mismatched marrow processing for transplantation: results of 114 procedures and analysis of immediate adverse events and hematopoietic recovery. Transfusion 2006;46:398-402. 29. Scholl S, Klink A, Mugge LO, et al. Safety and impact of donor-type red blood cell transfusion before allogeneic
evaluation of a new leukapheresis technology in non-
peripheral blood progenitor cell transplantation with major
cytokine-stimulated donors. Vox Sang 2014;106:248-55.
ABO mismatch. Transfusion 2005;45:1676-83.
1282 TRANSFUSION Volume 55, June 2015
