Abstract
Umbilical cord blood banks use two methods to store frozen umbilical cord blood (UCB): red cell reduction (RCR) or plasma depletion (PD). The RCR method centrifuges cord blood in hetastarch or albumin to isolate 21 ml of cord blood containing mostly white blood cells, adds 4 ml of 50% dimethyl sulfoxide (DMSO), and then freezes the resulting 25 ml of cell suspension. The PD method removes plasma, saves all the cells, and freezes the cells in 10% DMSO. PD UCB units are cheaper to process but more expensive to store and somewhat more troublesome to thaw. However, when properly thawed and washed, PD UCB units have as many or more total nucleated cells (TNCs), CD34+ cells, and colony-forming units (CFU) than RCR units. Two studies suggest that PD units have 20–25% more TNCs, MNCs, and CD34+ cells, as well as two to three times more CFU than RCR units. Higher TNC, CD34+, and CFU counts predict engraftment rate with faster neutrophil and platelet recovery. PD units have high engraftment rates with low mortality and high disease-free survival, comparable with clinical results of treatments with RCR units. One recent series of studies suggests that PD units are more effective for treating thalassemia with 2-year survival rates of 88%, disease-free survival rates of 74%, and 100% cure rate for children under age 7, compared to only 61% overall survival and 23% disease-free survival rate in thalassemic children treated with RCR units. These findings suggest that PD units not only have more TNCs, CD34+ cells, and CFU than RCR units but also have high engraftment rates and may be more effective for treating certain conditions such as β-thalassemia.
Keywords
Introduction
Two methods are currently used to store umbilical cord blood (UCB) units, that is, the red cell-reduced (RCR) and plasma-depleted (PD) methods. Developed in the early 1990s by Pablo Rubenstein of New York Blood Bank (22), the RCR method centrifuges the cord blood in a hyperosmotic solution (hetastarch or albumen), removing all but 21 ml of the cells and fluids, and adds 4 ml of dimethyl sulfoxide (DMSO). The plasma-depleted (PD) method, developed by Robert Chow (6) in the late 1990s, removes plasma but retains all the cells. PD units are typically 80–120 ml in volume compared to 25 ml of RCR units.
Cord blood banks add DMSO to cord blood to protect cells during freezing. DMSO reduces ice formation inside cells and allows >90% of cells to survive freezing. However, >1% DMSO is toxic to blood cells at body temperature (37°C). For that reason, care must be taken to minimize DMSO administration to patients, and DMSO is added to cord blood just before freezing and removed shortly after thawing so that the cells are not exposed to 1% for periods exceeding 30 min. If cord blood is exposed to >1% DMSO for 30 min or more, cord blood cells will die and clump together. This may result in emboli to the heart, chest pain, and other symptoms when cord blood is infused intravenously.
PD UCB units are more troublesome to thaw and wash, due to their larger and variable volume. However, when they are properly thawed and washed, PD UCB units have more total nucleated cells (TNCs), CD34+ cells, and colony-forming units (CFU) than RCR UCB units (5). Published engraftment rates for PD UCB units are similar to or better than RCR units (6,20). Direct comparisons of PD and RCR units for treatment of specific disorders, however, are rare. One series of studies suggests that PD UCB units are more effective for treating children with β-thalassemia (10) than RCR UCB units (23). These studies will be reviewed in this article.
RCR versus PD Cord Blood
In 1995, Rubinstein et al. (22) published a classic paper describing the red blood cell depletion (called red cell-reduced or RCR in this article) method that most cord blood banks use. Cord blood (usually 80–200 ml) was collected from the umbilical cord vein into a sterile bag containing an isotonic solution of citrate, phosphate, dextrose, and adenine. To minimize freezer storage space required, they eliminated most red blood cells by adding 6% hydroxyethyl starch (HES) until the HES concentration was 1.2% in the blood, centrifuging the blood to separate red and white blood cells, and transferring the latter into a smaller bag that was centrifuged again to remove excess plasma until the volume was 21 ml. They added 4 ml of 50% DMSO solution to reach a final volume of 25 ml and 10% DMSO concentration. The cells were then frozen in programmed stages and stored in liquid nitrogen at −196°C. Cord blood frozen at −196°C in 10% DMSO can be thawed as long as 20 years later with >80% viability of the cells (26).
The RCR method requires at least three bags, that is, a collection bag, a red cell waste bag, and a freeze bag, as well as a special centrifugation device that can detect and separate red blood cells from the white blood cells. Figure 1 illustrates the RCR method implemented on a device called Thermogenesis AXP (made by General Electric; Thermogenesis Corporation, Rancho Cordova, CA, USA). Instead of using hetastarch as the osmotic material for the gradient, this approach uses albumen as the osmotic material. The device centrifuges the cord blood until a gradient of cells develops. The cell suspension is then allowed to flow out a tube at the bottom of the bag into a red cell waste bag until the flow of cells falls below a certain criteria for redness detected by a laser. At that point, 21 ml of mostly white blood cells is gated to a freeze bag. Four milliliters of 50% DMSO is added to the 21 ml of white blood cell suspension to achieve 10% DMSO. This bag of 25 ml of mostly white blood cells is then frozen. The collection bag and the red cell bag are discarded.
The red cell reduction (RCR) and plasma depletion (PD) methods. In the RCR method, umbilical cord blood (UCB) is centrifuged, and red cells are dripped into a waste bag until a laser detects less redness and gates 21 ml of mostly white blood cells into a freeze bag. Then 4 ml of 50% dimethyl sulfoxide (DMSO) is added to the freeze bag to reach a final volume of 25 ml and 10% DMSO concentration. In the PD method on the right, the UCB is centrifuged to separate the cells and plasma. The plasma is squeezed into a separate plasma bag, and 50% DMSO is added until the final DMSO concentration is 10% in the cord blood. The cord blood is then transferred to a freeze bag.
Developed by Chow et al. (5) in the late 1990s, the PD method is simpler than the RCR method. As shown in Figure 1, the cord blood is centrifuged to pack the cells at the bottom of the bag, allowing removal of plasma from the top of the bag. Since the hematocrit of cord blood is typically about 50%, the method reduces the volume of the cord blood by 50% and saves all the cells in the blood. Since cord blood units are 80–200 ml in volume; this results in 40–100 ml of packed cells. The PD method only involves centrifugation of the collection bag and manual squeezing of the plasma from the top of the collection bag followed by transfer of the blood into a freeze bag. Fifty percent DMSO solution is added into the bag until the final DMSO concentration is 10% before freezing.
Several groups have optimized freezing and thawing procedures for cord blood. In 1996, Donaldson et al. (7) compared effects of different DMSO and HES concentrations on cluster of differentiation 34-positive (CD34+) cells (the cells thought to represent hematopoietic stem cells). They achieved 85.4 ± 28.4% recovery of CD34+ cells frozen at 1°C/min freezing rate in 5% DMSO. When they reduced DMSO to 2.5%, the CD34+ recovery fell precipitously to 12.2 ± 10.0%. Increasing DMSO to 10% and variations of HES had no effect on CD34+ recovery. In 2006, Meyer et al. (18) showed that 10% DMSO and 2% albumin were better for high cell concentrations (5 × 106 cells/ml) and that fast addition and removal of DMSO before and after freezing resulted in 89% and 88% recovery of total nucleated cells and CD34+ cells, respectively. In 2007, Woods et al. (28) found that adding and removing DMSO at ambient (room) temperature was better. Most groups (30) now cool the blood down to ambient or lower temperatures before adding DMSO to 10% concentration before freezing the cord blood at the rate of −1° per min.
Thawing RCR and PD Units
DMSO is toxic to cells exposed to concentrations of >1% at room or higher temperatures for longer than 30 min. Therefore, UCB must be diluted or washed to reduce DMSO concentrations shortly after thawing. RCR units are only 25 ml, so it is possible to dilute them in one step (e.g., adding 225 ml) to get DMSO concentrations down to 1% in one step or in two steps involving a dilution and wash. PD units have higher and variable volumes ranging from 80 to as much as 200 ml. One-step dilution of a PD unit would require 800–2,000 ml of fluid, and this might be too much volume to give to babies. Serial washes are necessary, that is, adding a volume of solution followed by centrifugation, removal of supernatant, and redilution, as illustrated in Figure 2. Jaing et al. (8–10) circumvented the need for washing the cord blood by thawing the cord blood unit at the bedside and diluting the cord blood intravenously as soon as it has thawed, but this should be done only by experienced clinicians and after careful consideration of the DMSO dose. Barker et al. (2) and Regan et al. (21) advocated a single-step dilution “no-wash” strategy for RCR units.
Serial washes and single dilution. The top diagram shows serial washes of a tube of cord blood. A volume of wash solution, equal to that of the cord blood, is added to the tube. This dilutes the DMSO concentration by half to 5.0%. The tube is centrifuged, the resulting 50% volume of supernatant is discarded, and an equal volume of wash solution is again added, reducing the DMSO concentration to 2.5%. The next wash halves the DMSO concentration to 1.25%. The bottom diagram shows a single dilution step by adding 7× the volume of cord blood with the wash solution, containing 2.5% dextran and 5% albumin in saline. In a single step, this reduces DMSO concentration to 1.25%.
Figure 3 illustrates one recommended approach to thawing RCR units (22). The procedure requires a transplant and a transfer bag in addition to the freeze bag. The first step is to add 12.5 ml of thaw solution to the freeze bag. As the frozen cells thaw, the second step is to drain the freezer bag into the transplant bag. The third step is to add 12.5 ml of thaw solution into the transplant bag and mix by passing the solution between the freeze and the transplant bag. The fourth step is to add 120 ml of thaw solution to the transplant bag, ending up with 170 ml and 1.5% DMSO. The fifth step is to centrifuge the transplant bag and to squeeze two thirds of the supernatant into the transfer bag, leaving about 50 ml of cord blood. The sixth step is to add 25–50 ml of thaw solution to the transplant bag and weigh the bag to ascertain an estimate of its volume and DMSO concentration. The seventh step is to remove samples for TNCs, CD34+ cell counts, CFU, and sterility tests. Finally, if cells are present in the transfer bag and needed, they can be recovered.
RCR thawing method. The freezing bag initially contains 25 ml of RCR cord blood and 10% DMSO. Add a 12.5-ml aliquot of thaw solution to the freeze bag, drain the thawed blood into the transplant bag, add 12.5-ml aliquot to the transplant bag and mix by passing the blood solution between the two bags. Then add 120 ml of thaw solution to the transplant bag, remove samples for analyses [CBC, count of blood cells; CFU, colony-forming units, cluster of differentiation 34-positive (CD34+) represent progenitor cells, and sterility tests for bacteria], centrifuge the transplant bag, squeeze two thirds of the supernatant to the transfer bag, add 25–50 ml of thaw solution to the transplant bag, and weigh the bag to estimate volume.
Thawing and washing of PD units is complicated by the larger and variable volumes of PD units. Frozen PD units range from 80 to 200 ml in volume. Figure 4 illustrates one procedure for thawing and washing PD units from a 300-ml freeze bag. The frozen PD unit is first placed inside a sterile plastic bag and placed in an ice bath for 5 min. The bag is removed when the blood is “slushy” and placed on ice packs. A wash solution contains 5% human serum albumen and 8% dextran 40 (a low molecular weight polysaccharide available commercially as Gentron 40, Hyskon, 10% LMD, or rheomacrodex) to balance the osmolarity of 10% DMSO. An amount of wash solution equivalent to 1.5× the unit volume is added to the freeze bag, and the contents of the freeze bag are squeezed into a transfer bag. An additional 1× volume of the unit volume is then added. At this point, the original 10% DMSO should have been reduced to about 2.86%. The contents of the transfer bag are transferred to 50-ml centrifuge tubes, spun in a refrigerated centrifuge (Sorvall HS-4) at 1,860 × g (3,080 rpm) for 15 min at 0–8°C with centrifuge brake set at medium. From each tube, all but 5 ml of the blood suspensions are aspirated; 5 ml of wash solution is added with a sterile pipette, mixed, and then additional wash solution is added until the volume is 50 ml. This dilutes the DMSO concentration 10-fold to about 0.3%.
A thaw and wash procedure for PD units. Place unit in a sterile plastic bag and put in an ice bath for 5 min. If the unit contains 100 ml, the initial dilution step should have resulted in 10% * 100/350 = 2.78% DMSO. The final wash and dilution step should reduce the DMSO concentration 10-fold to about 0.3%.
The above methods share three principles. First, the cord blood is kept cold (on ice) until the DMSO concentration is lowered to less than 1%. At no time is the blood allowed to be in contact with >1% DMSO at room temperatures or higher. Second, the cord blood is thawed by using a wash solution that is matched to the osmolarity of cord blood solutions containing 10% DMSO to minimize osmotic shock to the cells. Third, when diluting the blood suspensions, wash solution is always added to the cells. Cell suspensions are not added to the wash solution in case there is any osmotic difference between the cell suspension and the wash solution. Careful adherence to these practices should maximize viability of thawed UCB cells, whether processed by the RCR or PD method.
Comparisons of Total Nucleated and other Cell Counts
In 2011, Barker et al. (3) wrote in an article entitled “How I treat: The selection and acquisition of unrelated cord blood graft” that they usually discount the TNC counts of PD UCB units by 25% compared to RCR UCB units. They called PD units “RBC replete” units and claimed that such units have higher TNC counts because of the higher nucleated RBC and granulocyte content, given that no cells are removed. Barker et al. (3, p. 2334) claimed that many treatment centers correct the TNC dose of RBC replete units and that “this strategy is controversial, and the most appropriate correction factor is unknown.” Nevertheless, they stated, “Our current institutional policy is to use a correction factor of 0.75 (i.e., TNC × 0.75 = corrected TNC for RCB replete units).”
In 2011, Chow et al. (5) compared numbers of TNCs, CD34+, and other cells in cord blood units frozen with the PD and RCR methods, as shown in Figure 5. They took 10 fresh cord blood units and processed half of each unit using the PD and half using the RCR method. After freezing, the median recovery was 90% TNCs and 88% CD34+ cells. Comparisons of PD and RCR units indicated that the former have higher mean TNC counts of 124% and higher CD34+ counts of 121%. Percent changes of postthaw (post) to prefreeze (pre) TNC counts were −4.12 ± 2.11% and −21.43 ± 3.32%, respectively, for PD- and RCR-processed UCB, statistically different at p < 0.0003. The percent change between pre- and post-CD34+ counts were +1.13 ± 4.95% and −14.30 ± 6.27%, respectively, for PD and RCR UCB (p = 0.0692). While the percent changes of pre- and post-MNC were −7.39 ± 11.24% and −20.57 ± 14.88% for PD and RCR UCB, they were not statistically significant (p = 0.2075). Postthaw PD UCB also had higher CFU counts, that is, 225% CFU-GM (granulocyte and monocyte), 201% CFU-GEMM (granulocyte, erythrocyte, monocyte, megakaryocyte), 286% total CFU, and 187% VSEL (very small embryonic-like) cells, compared to CFU counts in RCR units.
Comparison of volume, total nucleated cell (TNC) count, mononuclear cells (MNCs), and CD34+ cells in 10 cord blood units divided equally and processed by the plasma depletion (PD) and red cell reduction (RCR) methods. The data were extracted from Chow et al. (5). The graphs show prefreeze (pre), postthaw (post), and percent change of UCB unit volumes (ml), total nuclear cell counts, mononuclear cell counts, and CD34+ cell counts. Plasma depletion causes significantly less TNCs and CD34+ cell loss than red cell reduction.
In 2010, Basford et al. (4) compared UCB processed by the PD method against the original Rubinstein method (density gradient with albumen), a modified Rubinstein method (Hetastarch), and two other methods called PrepaCyte CB (BioE, St. Paul, MN, USA) and Sepax (Biosafe, Geneva, Switzerland; automated centrifugation device). They found that thawed PD units had significantly higher CD45+, CD34+/CD133+, T-cell, B-cell, and CFUs than UCB prepared by the Rubinstein and modified Rubinstein methods and about the same as UCB prepared by the PrepaCyte CB and Sepax methods. These results confirm that PD UCB units have higher mean TNC, CD34+, and CFU counts than RCR units isolated and frozen from the same cord blood.
TNC counts are a good predictor of engraftment rate (16,27). However, Yoo et al. (29) found that while both TNC counts and CD34+ cells predicted speed of neutrophil engraftment, only CFU CM (granulocyte/macrophage) counts predicted the speed of platelet recovery. Total CFU counts are strong independent predictors of both neutrophil and platelet engraftment (19). Recent studies suggest that increasing the dose of cells through double unit UCB transplants (1,15,24) or coinfusion of adult CD34+ cells alongside UCB transplants markedly increase engraftment rate (25), while transplanting ex vivo expanded progenitors from the same cord blood unit (17) also improves engraftment rate.
UCB units with higher TNC, CD34+, and CFU counts are more likely to result in rapid engraftment and recovery of neutrophil and platelets. UCB PD units generally have higher TNC, higher CD34+, and higher CFU counts than UCB RCR units. This is not surprising since the PD method saves all the cells in the blood, while the RCR discards some of the white blood cells along with the red blood cells. Ultimately, however, the proof is in the pudding, that is, clinical efficacy in the form of improved survival and cure of the conditions being treated with the cord blood. In the past decade, over 2,000 units of PD UCB units have been transplanted in over 300 treatment centers around the world (personal communication, Stemcyte, Inc.).
Therapeutic Efficacy of PD and RCR Cord Blood Units
In 2007, Chow et al. (6) reported the clinical outcomes of 118 mostly pediatric patients who received plasma-depleted cord blood units with mean and median nucleated cell doses of 7.6 × 107 and 5.6 × 107 cells/kg, respectively. The median times for engraftment were 22 and 50 days, respectively, for absolute neutrophil count (ANC = 500) and platelet (TPC = 20K) recovery; 90 ± 3% and 77 ± 5% of the patients showed neutrophil and platelet recovery. Acute (grades III–IV) and chronic graft-versus-host disease (GVHD) rates were 13 ± 4% and 17 ± 6%, respectively. Treatment-related mortality was 16 ± 3%. One-year survival and relapse-free rates were 65 ± 5% and 51 ± 6%. These represent unselected patients with many different hematopoietic and malignant conditions, making comparisons with other studies (15,16,27) difficult. However, the observation that 90% and 77% of these patients recovered neutrophil and platelet function suggests a large majority of PD units engrafted.
In 2012, Petz et al. (20) reported an analysis of 120 pediatric patients with nonmalignant disorders treated with PD units. The children were treated in 29 US and 17 international centers, ranged from 0.1 to 14 years (3.5 years median), 4 to 61 kg (15 kg median), and 58% were male. In terms of human leukocyte antigen (HLA) matching, 22% were 6:6 matched, 40% were 5:6 matched, 39% were 4:6 matched, and 5% were less than 4:6 matched. Median prefreeze TNC dose was 10.5 × 107/kg and CD34+ dose was 3.7 × 105/kg. Only 24% of the patients did not receive myeloablative regimens. The median times for myeloid and platelet engraftments were 21 and 49 days, respectively. Cumulative incidence of acute graft-versushost disease (aGVHD) was 38 ± 5%, and 19 ± 4% had severe grades III–IV. The Kaplan–Meier estimates of 3-year transplanted related mortality, overall survival, and disease-free survivals were 20 ± 4%, 79 ± 4%, and 70 ± 6%. These data are comparable or better than results from RCR transplants. However, because outcomes depend on diseases treated with UCB transplants, it is important to compare clinical results for matched disorders.
Ruggeri et al. (23) and Jaing et al. (8,10–14), respectively, described outcomes of RCR and PD UCB transplants to replace bone marrow in children with β-thalassemia. A multicenter study based on cases from the Eurocord Registry, the National Cord Blood Program (in the US), the New York Blood Center, and the Center for International Blood and Marrow Transplant Registry, the Ruggeri study (23) described the outcomes of single RCR cord blood transplants in 35 children with thalassemia and 16 children with sickle cell anemia. Only 24 of 51 (47%) patients recovered neutrophil counts. The median dose of cells transfused was 5 × 107 cells/kg (range 2.4–23). The 2-year probability of disease-free survival was 45% in 35 thalassemic patients that received a cell dose of >5 × 107 cells/kg and an abysmal 13% with patients that received lower total nucleated cell doses. The overall survival and disease-free survival of thalassemic children in the Ruggeri study were 62% and 21%, respectively. The authors concluded that primary graft failure was the predominant cause of treatment failure in 20 patients with thalassemia and suggested that only units with cell doses greater than 5 × 107 cells/kg should be used to treat thalassemia.
Jaing et al. (8,10) described 35 children with thalassemia treated with 40 PD cord blood units. The 5-year overall and disease-free survivals were 88.3% and 73.9%, respectively. At 2 years, 30 of 35 patients (85%) were alive and transfusion independent. The cumulative incidence of treatment-related mortality was 11.7% at 5 years. Fourteen (40%) patients developed chronic GVHD that resolved over time. Moreover, if one considers only children under 7, none died and all were cured, that is, transfusion independent. Comparison of the two studies revealed no significant differences in HLA mismatches or age differences between the Ruggeri and Jaing studies. The only differences were that Jaing et al. used only PD units, five children in the Jaing study received two cord blood units, and the children received higher median nucleated and CD34+ cell doses, respectively, 7.8 × 107 cells/kg (range 2.8–14.7) and 4.0 × 105 cells/kg (range 1.7–19.9), than in the Ruggeri study.
Conclusions
Most UCB banks use either the RCR method or the PD method to process cord blood for cryopreservation.
The RCR method centrifuges the cord blood and discards all but 21 ml of the blood before adding 4 ml of 50% DMSO and freezing the cells. The RCR method results in 25 ml of mostly white blood cells.
The PD method centrifuges the cord blood and discards the plasma while retaining all the cells. DMSO is then added to the bag, and the cord blood is frozen. PD units are typically two to three times bigger than RCR units.
DMSO is toxic to UCB cells at concentrations of >1% for longer than 30 min. UCB frozen in 10% DMSO must be washed or diluted to less than 1% DMSO concentration shortly after thaw.
Cord blood should be kept cold during dilution so that it is not exposed to >1% DMSO at ambient or higher temperatures. Wash solutions should include 5–8% dextran and 2% albumen to reduce osmotic shock as the DMSO is removed.
PD UCB units have 20–25% TNCs and CD34+ cells, as well as two to three times more CFUs than RCR units. These measures usually correlate with higher probability and rates of engraftment.
PD units have been transfused into over 2,000 patients with malignant and nonmalignant hematological disorders, showing high engraftment rates with 88% neurotrophil and 90% platelet recovery.
In children with β-thalassemia, PD units appear to be much more effective than RCR units, the former yielding overall and disease-free survival of 88% and 74%, compared with 62% and 21% for RCR units.
PD units take up more freezer space and are more troublesome to thaw, but they have more TNCs, CD34+ cells, and CFUs than RCR units. PD units seem to be as or more effective for replacing bone marrow, particularly for β-thalassemia.
Footnotes
Acknowledgments
Dr. Young is the Global Medical Director of Stemcyte. This is a voluntary position, that is, he receives no financial compensation, stocks, or other payments for doing this work. Stemcyte supports research that he does, including donations of umbilical cord blood for clinical trials of spinal cord injury and laboratory research that he does on umbilical cord blood cells at Rutgers University.