wbc reduction in cryopreserved rbc units

B L O O D C O M P O N E N T S

WBC reduction in cryopreserved RBC units

Francoise G. Arnaud and Harold T. Meryman

BACKGROUND: WBC reduction of blood componentsby filtration is widely practiced to decrease the inci-dence of alloimmunization. Freezing RBCs reduces theWBC load but is insufficient to achieve the currently rec-ommended US limit of 5 × 106 cells per unit.STUDY DESIGN AND METHODS: Blood units wereWBC reduced by filtration or by buffy-coat (BC) removaland then frozen in the presence of a high-glycerol con-centration. The count of residual WBCs was determinedby flow cytometry after deglycerolization.RESULTS: Without WBC reduction, the total number ofWBCs present after freezing and thawing was 11.5 ±9.2 × 106 WBCs per unit (n = 18). Particulate residuesfrom monocytes and neutrophils that were detected inthe remaining cell populations were positive for CD66b,CD3, CD14, and CD41. Removal of 40 mL of BC at thetime of blood collection lowered the number of WBCsafter freezing and deglycerolization to 1.9 ± 1.20 × 106

per unit (n = 11). Similar results were obtained whenonly 20 mL of BC was removed using a modified blood-bag design. Unfiltered RBC units that were stored for 15days at 4°C after BC removal contained fewer than 5 ×106 WBCs after deglycerolization. Units WBC reducedby filtration before freezing had no detectable WBCsafter thawing and washing (n = 14) and did not containparticulate residues. Filtration after deglycerolizationwas effective in reducing the WBC count below 106,although some debris was still present.CONCLUSION: RBC freezing alone will not reduce re-sidual counts to recommended levels. However, initialremoval of BC can provide an economical alternative toWBC filtration for cryopreserved units. Units that werenot WBC reduced before freezing can be filtered afterdeglycerolization when needed.

It is now recognized that the presence of WBCs dur-ing transfusions of RBCs can induce alloimmuniza-tion,1,2 and WBCs and platelets reacting with recipi-ent antibodies cause nonhemolytic febrile

transfusion reactions.1,3,4 WBC reduction has beenshown to decrease the frequency of primary HLA alloim-munization among patients with hematologic malig-nancy5,6 and to decrease the frequency of recurrent fe-brile nonhematologic transfusion reactions.7,8 In the US,blood units with a residual WBC count below 5 � 106

qualify as WBC reduced by the FDA,9 and in some Euro-pean countries, guidelines define acceptable WBC reduc-tion as 1 � 106 WBCs per unit or less. Although WBCreduction reduces the incidence of febrile reactions,studies have documented that it does not eliminate suchreactions in all recipients.10

Cryopreservation, which enables the long-term stor-age of RBCs, is designed to maximize RBC recovery. How-ever, during this process, significant numbers of WBCsare lost because the freezing procedure is not optimal forthese cells.11-13 In 1980, Meryman et al.,14 using a deglyc-erolizing device (Haemonetics Model 102, Braintree,MA), reported a reduction to 30 � 10 � 106 WBCs perunit. More recently, Valeri et al.,15 using a newer device(Haemonetics Model 215), reported depletion to 10 � 8� 106 WBCs per unit. Neither of these freezing proce-

ABBREVIATION: BC = buffy coat.

From the Naval Medical Research Center, Combat Casualty

Care, Resuscitative Medicine Department, Silver Spring, Mary-

land.

Address reprint requests to: Francoise Arnaud, PhD, Naval

Medical Research Center, Combat Casualty Care, Resuscitative

Medicine Program, 503 Robert Grant Ave, Silver Spring, MD

20910-7500; e-mail: [email protected].

This Research was supported in part by ONR grant

(#0603707 N00542.1WP.A0069). Statements and opinions ex-

pressed in this manuscript represent those of the authors and

are not to be construed as official or reflecting the views of the

Department of the Navy.

Received for publication July 26, 2002; revision received

November 1, 2002, and accepted November 25, 2002.

TRANSFUSION 2003;43:517-525.

Volume 43, April 2003 TRANSFUSION 517

dures produced RBC units that can qualify as WBCreduced.

WBC reduction using commercially available filterscan surpass the US standard, producing a WBC reductiongreater than 3 logs,9,16-18 and WBC reduction of trans-fused blood components has become increasingly com-mon, although not universally applied in all centers.10

The use of WBC filters before RBC cryopreservation re-sults in the loss of approximately 20 mL of blood volume,adds supplemental expenses, and requires additionaltime and manipulation.16,19 Therefore, we examined al-ternative methods for WBC reduction and the goal of thisstudy was to evaluate the extent of WBC reductionachieved in frozen/thawed RBC units with and withoutprior buffy-coat (BC) removal. Glycerolization and de-glycerolization of RBCs was performed using an appara-tus (Mission Medical M1000, Fremont, CA) that uses ahollow fiber separator instead of centrifugal separationfor processing RBCs.20

MATERIALS AND METHODSFreezing and thawingAll blood units used for this study were collected accord-ing to the FDA guidelines for informed consent. The pro-tocol was approved by the Human Subject Use Com-mittee of the Naval Medical Research Center (IRB# 31514).

Blood was collected from healthy volunteers at localBlood Centers and delivered either as whole blood or aspacked cells reconstituted with AS-3 or AS-5. Wholeblood was centrifuged at 2000 � g for 6 minutes (J6-HCcentrifuge, Beckman, Fullerton, CA), plasma was re-moved, and 90 mL of either AS-3 or AS-5 was added to theRBCs. In some experiments, removal of BC was per-formed using an inverted V bag (� bag). The upper half ofthe bag was narrowed to an inverted V with an angle ofapproximately 55� at the top so that the BC could migrateto the orifice of the bag more effectively. All RBC unitswere stored at 4�C either for 6 to 7 or for 15 to 16 daysbefore freezing. At that time they were glycerolized usinga fully automated and closed apparatus that conductsboth glycerolizing and deglycerolizing (M1000, MissionMedical).20 The general procedure was comparable to thecurrently licensed protocols developed by Meryman andHornblower21 and by Valeri.22 A buffered solution of 6.2M glycerol (Glycerolyte, Baxter, Deerfield, IL) was addedto the cells at a fixed ratio of 1.3 (vol/wt), yielding a finalglycerol concentration of approximately 3.8 M (35% [wt/vol]).22 The glycerolized RBC unit was centrifuged at 2000� g for 6 minutes and the supernatant removed. A layerof agglutinated WBCs and platelets was removed in unitsnot previously WBC reduced. The glycerolized RBCs werethen frozen in a mechanical freezer (Revco, Ashville, NC)at –80�C.

The frozen cells were thawed in a dry-heat warmingsystem (Model 3614, Forma, Marietta, OH) at 37�C forapproximately 15 minutes and then processed in theM1000 for glycerol removal. A solution of 12-percentNaCl (Baxter, Deerfield, IL) was added to the thawed cellsat a fixed weight ratio of one part saline to 6.5 parts RBCsfollowed by a 3-minute equilibration period. The washsolution consisted of 1700 mL of isotonic glucose-salinesolution (0.9% NaCl to 0.2% glucose, Baxter). The RBCswere repeatedly cycled through the hollow fiber unit thatseparated supernatant solution (glycerol and free Hb)from the cells. The final washed product had a Hct rang-ing from 50 to 65 percent. Freezing, thawing, and wash-ing of all units were performed under sterile conditionsand asepsis was maintained subsequently.

WBC depletionSeven depletion protocols were tested.

Group 1. No initial BC removal before glyceroliza-tion. Freezing and thawing take place after 7 days of stor-age at 4�C.

Group 2. Same procedure as Group 1 after 15 days ofstorage at 4�C.

Group 3. No initial BC removal. After 7 days of stor-age at 4�C, the RBCs were glycerolized and excess glycerolsolution removed together with 10 mL of BC.

Group 4. 40 mL of BC was removed at the time ofcollection during component separation. Freezing andthawing take place after 7 days of storage at 4�C.

Group 5. Same procedure as Group 4 with 20 mL BCremoval using the modified bag (� bag).

Group 6. Fresh units were WBC reduced at the timeof component separation using BPF4 WBC filters (PallBiomedical, NY) and were frozen after 7 days of storage at4�C.

Group 7. Same procedure as Group 1 except that thethawed units were filtered with BPF4 filters after the finalwash.

WBC measurementThe initial WBC count was determined for all units. WBCsin unfrozen and thawed/deglycerolized samples were de-termined by electronic, manual, or flow cytometrycounts. Electronic counting (ABX Pentra-60, ABX, Mont-pellier, France) has a high threshold (0.1 � 106 WBCs/mL) and was used for counting unfrozen samples.Manual counts were performed with an improved Neu-bauer counting chamber (Hausser Scientific, Horsham,PA) for samples with a concentration as low as 1 � 103

WBCs per mL. A total of 100 �L of blood was diluted witha lytic solution (8.27 g NH4Cl, 0.84 g NaHCO3, 37.2 mgEDTA-Na/L with a final pH of 7.4 and an osmolality of300 mOsm/kg) and mixed for 15 minutes on a shaker.This iso-osmotic solution has been widely used for WBC

ARNAUD AND MERYMAN

518 TRANSFUSION Volume 43, April 2003

enumeration and does not induce morphologic cell dam-age.2,23 The sample was then spun at 2000 � g for 8minutes (Eppendorf Centrifuge, Model 5415c, Westbury,NY), and the supernatant was discarded leaving 50 �L ofresidual volume on the pellet, which was then resus-pended with 50 �L of Turk’s solution (0.01% Violet Gen-tian). After 15 minutes of staining, 10 �L of the samplewas examined using a 40� objective (Nikon LaboPhot,Image System, Baltimore, MD). Cellular elements with adefined cellular structure (i.e., maintenance of the shapeand diameter and presence of nuclear entity) werecounted, however, this did not presume of their function.

WBCs were also counted by flow cytometry (Leuko-Count kit, Becton Dickinson, San Jose, CA). Ranges of 1 to300 � 103 WBCs per mL could be detected (from BectonDickinson Manual recommendations). A total of 400 �Lof the lytic-detergent solution was added to 100 �L ofblood. The tube was gently mixed and read on a flowcytometer (FACScan, Becton Dickinson, San Jose, CA).The reagent of the kit contained buffer and detergent topermeabilize WBCs and PI to label the nuclear DNA ofWBC. RNAse was used to digest residual cellular RNA thatcould interfere with the PI labeling. WBCs were countedin reference to a known number of calibrated and taggedbeads present in the TruCount tubes.

WBC phenotypesPhenotypes of residual WBCs were also examined usingMoAbs (Pharmingen, San Diego, CA) for specific surfaceepitopes. WBCs were identified with PE-labeled anti-CD45. Differentiation between T-cells, B-cells, mono-cytes, granulocytes, and platelets was achieved by FITC-labeled anti-CD3, -CD19, -CD14, -CD66b, and -CD41,respectively. Unfrozen whole-blood samples were diluted10� with PBS, whereas thawed and washed bloodsamples were used undiluted. A volume of 100 �L of cellsample was dual labeled, and after 20 minutes incuba-tion, 400 �L of freshly prepared lytic solution was added.The data were analyzed on the flow cytometer (FACScan).

Screening for apoptosisThe WBCs remaining after deglycerolization were con-centrated with a Ficoll hypaque density gradient (10 mLblood for 5 mL Ficoll). After washing in PBS, the cellswere layered on glass slides by cytospin and fixed in para-formaldehyde (Sigma, St Louis, MO). After membrane di-gestion with Proteinase K (Roche Molecular Systems,Pleasanton, CA) the cells were screened for DNA frag-mentation by Tunel assay using an ApopTag kit (Inter-gen, Purchase, NY).

Microscopy was performed both on unfrozensamples and after freezing. The cells were embedded in amixture of Glutarahdehyde and paraformaldehyde(Sigma, St Louis, MO) and processed for electron micros-copy.

StatisticsWBC counts were subjected to statistical analysis usingpaired and unpaired Student’s t-test, linear regression,and CV. CV was calculated as the SD divided by the meanfor serially diluted samples. The results were consideredsignificant when the p value was less than 0.05.

RESULTSUnits frozen and processed in our laboratory with theM1000 had the following characteristics: unfrozen bloodunits reconstituted with AS (AS-3 or AS-5) after compo-nent separation had a volume of 282 � 56 mL and a Hctof 62.8 � 3.5 percent (n = 78). After deglycerolization inthe M1000, the volume was 304 � 48 mL with a Hct of55.4 � 2.7 percent (n = 78).

Fig. 1. (A) Comparison of WBC counting methods before

freezing. Correlation between electronic (�), flow cytometry

(�), and manual WBC counting (x axis). Regression line ob-

tained by least-square method (––– for electronic and ------

for flow cytometry). (B) Comparison of WBC counting meth-

ods after thawing and deglycerolization. Correlation between

flow cytometry and manual WBC counting methods (�). Re-

gression line obtained by least-square method (––––).

WBC REDUCTION IN BLOOD UNITS


Validation of the measurementsSamples were serially diluted over a wide range andcounted using the several methods of measurement de-scribed and their accuracy and reproducibility were ac-cepted when the CV was less than 25 percent. WBCcounts tested ranged from 0.5 to 10 000 � 103 WBCs permL before glycerolizing and from 0.5 to 200 � 103 WBCsper mL after deglycerolization. WBCs in unfrozensamples counted with the ABX apparatus served as base-line counts. Sample concentrations for manual countswere validated between 2 and 1000 � 103 WBCs per mLto give a CV of 21.5 percent or lower for nonfrozensamples, and a CV of 23.8 percent for deglycerolizedsamples. With flow cytometry, the validation range wasfrom 1 to 1000 � 103 WBCs per mL with a CV lower orequal to 14.6 percent for nonfrozen samples and 13.1percent for deglycerolized samples. Correlations betweencounting methods were established within the range ofreproducibility of the reading, and all three methods werein good agreement with each other (R2 > 0.86), with themanual methods being the least sensitive (Fig. 1A and B).No significant difference was found between the manualand flow cytometry methods for paired samples at con-centrations above 4 � 103 per mL. Below this concentra-tion, the cells were either not detected by manual countsor their counts presented a greater variability. Flow cy-tometry offered more reproducibility and was used forthe data presented. Electronic counting was not suitablefor cryopreserved samples.

WBC reduction in the different groupsTable 1 shows the total number and distribution of non-frozen WBCs after component separation, or after initial

removal of WBCs by different techniques, after periods ofstorage at 4�C as indicated. After 1 week of storage, thetotal WBC count was reduced by about one-third primar-ily due to the loss of neutrophils (p < 0.001), which areknown to undergo apoptosis after blood collection.2 Afteran additional week of storage, there was a further loss ofinitial WBCs. There was a similar significant decrease inlymphocyte and monocyte counts but not of neutrophils.When 40 mL of BC was removed at the time of compo-nent separation (Day 0), the WBC count in fresh bloodwas reduced by half (p < 0.001), with lymphocytes andmonocytes preferentially removed (p < 0.001). An addi-tional week of storage resulted in the loss of more neu-trophils but not lymphocytes or monocytes. A similarpattern of reduction was observed using the � bag. Therewas no significant difference in the WBC count or distri-bution after the removal of 40 mL of BC or when 20 mL ofBC was removed using the � bag. Filtration before freez-ing reduced all classes of cells by more than three logs.The SD of the results reflects the large discrepancy ininitial WBC counts due to differences among donors.

Table 2 illustrates the reduction in total WBCs andthe loss of RBCs after freezing, thawing, and deglycer-olization using the different protocols. Units stored for 7days when no WBC reduction was performed before glyc-erolization did not meet the FDA requirements, and WBCcounts were similar to those previously reported.14,15 Ad-equate WBC reduction was achieved only if the units hadbeen stored at 4�C for 15 days (p < 0.001, Group 2). Theremoval of 40 mL of BC (Group 4) resulted in satisfactoryWBC reduction after 7 days of storage (p < 0.001) but atthe cost of considerable RBC loss. BC removal in themodified bag was equally effective and reduced subse-quently the loss of RBCs (Group 5). The WBC count was

TABLE 1. Distribution of Leukocytes in blood prior to freezingStorage

time at 4°C(Days)

Total WBC109 per unit

Neutrophils109 per unit

(%)

Lymphocytes109 per unit

(%)

Monocytes109 per unit

(%) N

No leukodepletion 0 2.6 ± 0.6 1.75 ± 0.56 0.71 ± 0.14 0.18 ± 0.08 18(Whole blood) (63.5 ± 7.4%) (26.9 ± 6.9%) (6.6 ± 2.4%)

No leukodepletion 6.4 ± 0.6 2.0 ± 0.7 0.71 ± 0.45* 1.07 ± 0.46 0.09 ± 0.04* 18(35.8 ± 16.8%) (54.3 ± 18.3%) (4.5 ± 1.1%)

No leukodepletion 16.3 ± 4.2 0.9 ± 0.6 0.33 ± 0.25 0.46 ± 0.04* 0.03 ± 0.01* 9(35.1 ± 17.8%) (56.3 ± 14.8%) (4.0 ± 1.9%)

Removal of 40 ml 1.3 ± 0.5 0.97 ± 0.44* 0.1 ± 0.07* 0.03 ± 0.02* 11BC at Day 0 0 (83.4 ± 5.5%) (8.2 ± 4.1%) (2.4 ± 0.9%)

Removal of 40 ml 6.2 ± 0.4 1.0 ± 0.6 0.65 ± 0.27** 0.33 ± 0.31** 0.03 ± 0.02 11BC at Day 0 (62.8 ± 20.4%) (24.3 ± 16.5%) (2.8 ± 1.4%)

Removal of 20 ml 1.26 ± 0.26 0.99 ± 0.30* 0.14 ± 0.10* 0.04 ± 0.02* 6BC in modified � 0 (77.3 ± 13.9%) (11.7 ± 8.7%) (3.6 ± 2.1%)bag at Day 0

Removal of 20 ml 5.9 ± 0.5 1.15 ± 0.26 0.81 ± 0.23 0.18 ± 0.08 0.04 ± 0.02 6BC in modified � (68.1 ± 11.0%) (15.8 ± 8.0%) (3.3 ± 1.4%)bag at Day 0

Leukofiltered at 6.5 ± 0.7 <1 × 106 N/A N/A N/A 14Day 0

ARNAUD AND MERYMAN


below the minimum standard when units were filteredeither at the time of component separation or after de-glycerolization (Groups 6 and 7).

Scatter graphs (Fig. 2A-D) obtained by flow cytom-etry revealed the WBC population in the gated area R1(positive to PI; FL2) in comparison to beads gated in R2.There were other tagged particles present above R1 thatmay represent adhesion to other cells or debris (Fig. 2Aand B). Larger quantities of untagged particles below R1were detected after freezing and deglycerolization (Fig.2C). Because these untagged particles were largely absent inunits filtered either before or after freezing, they were prob-ably WBC debris or RBC fragments resulting from freezingdamage that were removed by WBC filtration (Fig. 2D).

An analysis of WBC phenotypes of fresh preparationsrevealed three distinct populations that consisted of es-sentially pure lymphocytes (CD3+) in gate R3, monocytes(CD14+) in gate R4, and granulocytes (CD66b+) in gate R5(Fig. 3). This distribution also matched the one obtainedby electronic determination (Table 1). After freezing anddeglycerolizing (Fig. 3C), there was still evidence of these

antigens, but the distribution was notcomparable to that seen in fresh blood(Fig. 3A and Table 3). Gate R5, for ex-ample, which represented neutrophilsin fresh blood, now displays a variety ofmarkers (CD3, CD14, CD41), probablythe result of antigens adhering to cells.This population was also found to ex-press a high nonspecific binding. TheR4 gate, which was specific for mono-cytes, is now contaminated with cellsexpressing CD3. After freezing, the ma-jority of residual WBCs after freezingwere CD3+, marking them as T-

lymphocytes. B-cells (CD19+) cells were not found afterfreezing. Granulocytes are notoriously sensitive both tofreezing and to 4�C storage so that cells displaying CD66bare most likely carrying adherent granulocyte debris. Ex-pression of multiple antigens on cells suggests that eachgate now includes fragments of cells adherent to othercells, making positive identification of surviving WBCsdifficult or impossible. Platelet antigen (CD41+) was alsofound in the mix of detectable phenotypes. Platelets areinvolved in the adhesion process and might have been afactor in the adherence of membrane fragments to intactcells. Platelet antigens were also found adhering tomonocytes in fresh preparations.

With blood units stored for 6 days at 4�C, a correla-tion (R2 = 0.623, p < 0.01) was seen between the prefreezeand postwash WBC counts (Fig. 4A) and, as shown inFigure 3, it is not possible to identify the relative sensi-tivity of individual phenotypes to freezing injury. No cor-relation was found between the residual WBC count andthe amount of RBC loss resulting from the freeze-thaw-wash procedure (Fig. 4B).

TABLE 2. Frozen/thawed/washed blood units

ProcedureWBC perunit × 106

RBC loss dueto depletion N

Group 1: 7-day storage 11.5 ± 9.2 0 18Group 2: 15-day storage 2.6 ± 1.3* 0 9Group 3: 10 ml BC removal after glycerolization

after 7-day storage9.5 ± 9.4 1.7 ± 0.8 ml 5

Group 4: 40 ml BC removal from standard bagand 7-day storage

1.9 ± 1.2* 30 ± 4.5 ml 11

Group 5: 20 ml BC removal from � bag and7-day storage

2.7 ± 0.8** 4.9 ± 1.7 ml 6

Group 6: Leuko-filtered and 7-day storage 0.03 ± 0.03 ∼10 ml 7Group 7: Leuko-filtered after deglycerolization 0.2 ± 0.2 ∼10 ml 8

Fig. 2. Flow cytometry scatter graph using the LeukoCount kit. (A) Fresh blood Day 0. (B) Blood stored at 4�C for 6 days.

(C) Frozen-thawed washed unfiltered blood. (D) WBC-depleted blood. The isolated WBC population (FL2 positive) is located

in gate R1. Gate R2 represents the beads. The population in the lower-left part of the graph is negative for both fluorochromes

FL1 and FL2 and consists of untagged debris. The FL2-positive population adjacent to R1 could be attributed to cells with

adherent debris.



After deglycerolization, residual lymphocytes viewedin the hemocytometer appeared somewhat abnormal,raising the question whether they were still viable, espe-

cially because high concentrations of glycerol are notsuitable for the cryopreservation of purified WBC subsets.Evidence of apoptotic cells was also found after 7 days of

storage (Fig. 5) because the Tunel assayindicated the presence of cells withfragmented DNA. After deglyceroliza-tion, the Tunel assay is difficult to in-terpret due to the high-fluorescencebackground. However, electron micro-graphs of the BC layer indicated thepresence of necrotic lymphocytes afterdeglycerolization (Fig. 5).

DISCUSSIONThe goal of this study was to determinewhether freezing of RBCs with or with-out additional WBC reduction couldachieve the current US standard of 5 �

106 residual WBCs per unit. This targetcannot be attained with RBC freezingalone or with a deeper removal of theinterface (10 mL of BC) after glycer-olization and before freezing. Counts of

TABLE 3. Phenotype expression of residual CD45+ WBCsA Distribution of WBCs in unfrozen blood (%)

Gate R3: Neutrophils R4: Monocytes R5: Lymphocytes N

Fresh blood CD45+ 71.2 ± 7.7 5.7 ± 0.2 23.1 ± 7.9 3Blood after BC

removal andstorage at 4°C.

CD45+ 90.3 ± 5.7 2.7 ± 0.4 7.0 ± 5.5 5

B Distribution of WBCs in frozen and deglycerolized blood (%)

Gate R3 R4 R5 N

Frozen anddeglycerolizedblood

CD45+ 19.4 ± 8.3 63.9 ± 14.4 17.6 ± 15.2 20

Distribution ofepitopes

R3 R5 R4CD3+ 66.4 ± 31.7 48.8 ± 29.7 50.3 ± 39.7 20CD14+ 6.14 ± 21.6 5.8 ± 9.0 8.8 ± 19.1 20CD66+ 68.7 ± 23.2 30.1 ± 29.0 5.9 ± 12.8 20

C. Evidence of Antigen of residual CD45+ WBCs in washed blood (%)

CD41+ CD66+ CD14+ CD3+ N

Frozen anddeglycerolizedblood

21.9 ± 15.1 18.5 ± 20.5 4.1 ± 5.8 37.1 ± 20 20

Fig. 3. Flow cytometry scatter graphs for fresh RBCs (A1, A2), RBCs stored 7 days at 4�C (B1, B2), and frozen deglycerolized

RBCs (C1, C2). Graphs A1, B1, and C1 are forward versus side-scatter plots. Graphs A2, B2, and C2 illustrate side-scatter versus

fluorescence. After labeling of the cells with CD45-PE MoAb, populations of lymphocytes (gate R3), monocytes (gate R4), and

granulocytes (gate R5) were identified in fresh blood. The markings indicate a general shift from the original gate after storage

and freeze-thaw and contribute to the general mixing of cells and debris.

ARNAUD AND MERYMAN


less than 5 � 106 WBCs are obtained if40 mL of BC is removed from freshunits at the time of component separa-tion. Although inexpensive, this has thedisadvantage of losing an average of 30mL of RBCs per unit. However, accept-able WBC depletion with the loss of lessthan 5 mL of RBCs is possible with theremoval of 20 mL of BC when the top ofthe collection bag is modified into aninverted V shape. WBC reduction bythese procedures requires that theRBCs be designated for freezing beforecomponent separation. Our data do notaddress the use of fresh RBC units fro-zen immediately after collection; allunits were frozen after 7 or 15 days ofstorage at 4�C. Adequate WBC reduc-

tion is also obtained with non-WBC-reduced RBCs storedfor 15 days before freezing. However, current regulationsrequire that RBCs be stored no longer than 6 days beforefreezing, although this limitation is arbitrary, has nophysiologic basis, and could be extended if warranted.The use of the BPF4 filters either before or after freezingprovided excellent WBC reduction, although when filtra-tion was performed after deglycerolizing, some WBCscould still be found. It could be that minute amounts ofremaining glycerol prevent full adhesion of WBCs ontothe fibers.

After freezing and deglycerolization, residual WBCsare essentially lymphocytes, confirming previous re-ports.11,19 All of the WBCs are not enclosed in the respec-tive gates set for fresh cells as indicated by the phenotypeanalysis, but WBCs are associated with adherent frag-ments and express more than one phenotype. The pres-ence of debris was also noticed with unlabeled cells orcells with a weakened expression of nuclear content, im-plying important structural damage. The majority ofWBCs become apoptotic during 6 to 7 days of 4�C stor-age.2 These degenerating cells are presumably respon-sible for the debris and cell fragments seen in flow cy-tometry after freezing. We have observed that cells afterdeglycerolization are mainly necrotic (Fig. 5) because thestructureless condensed nuclei (pyknosis) in some ofthese cells and the rupture of the membrane in otherssuggests irreversible cell injury, implying no function invivo.2 The release of cell by-products, including cytokinesfrom necrotic WBCs after freezing, raises concerns re-garding the inflammatory and febrile reactions they cancause,5-7,11,24 and it has been reported that nuclear pro-teins released from apoptotic cells in blood units mayinduce an immune response in multiple transfused pa-

Fig. 5. Microscopy. (A) WBCs from whole blood after TUNEL assay. (A1) WBCs

from fresh blood. Red PI counter stain indicates no apoptosis (20� magnifica-

tion). (A2) WBCs from blood stored for 6 days at 4�C. Apoptotic cells fluoresce

green with fluorescein (10� magnification). (A3) WBCs from deglycerolized blood.

The high background suggests a bleed of fluorescein from damaged WBCs. The

arrow indicates the remnants of a WBC (20� magnification). Red indicates PI

counterstain. Green indicates fragmented DNA labeled with fluorescein. (B) Elec-

tron microscopy of a damaged WBC in the BC layer after deglycerolization. Bar =

1 �m.

Fig. 4. (A) Relationship between WBC counts before and af-

ter freezing (�). A regression of the second order (R2 =

0.623) was obtained by Excel best fit. (B) Relationship be-

tween the total hemolysis of RBCs after thawing and deglyc-

erolization and the number of WBCs found after deglycer-

olization (�). Units were stored for 6 days at 4�C before

freezing.



tients.25 These factors are probably removed during thedeglycerolization process,26 and the cell remains willlikely be engulfed by recipient macrophages, reducingthe risk of an immune response.2

The presence of multiple epitopes on residual cellscould increase the risk of recipient alloimmunization,and, therefore, removal of cellular or nuclear antigenicfragments may be desirable and may argue in favor ofinitial WBC filtration. Filtration in conjunction with sub-sequent freezing, achieves nearly total WBC depletion. Ithas the added virtue of eliminating the cell debris andother adherent material on cells that are seen after freez-ing of unfiltered blood units. However, the incidence ofnonhemolytic febrile transfusion reactions is less than 1percent in all transfusions, and this rate does not differsignificantly after transfusions of WBC-reduced units.9,10

Thus, WBC reduction before storage does not totallyeliminate the immunomodulatory agents present instored RBC units, but subsequent washing does appear todo so.26 Washing is an integral part of deglycerolization,which can bring the WBC count below the 5 � 108 WBCthreshold that is effective in preventing most transfusionreactions.9 The question of cost is particularly relevant inthe case of cryopreserved blood units that are alreadyexpensive and time consuming. Do all cryopreservedblood units need to be WBC filtered if WBCs can be re-duced to 107 or below 5 � 106 with inexpensive BCremoval?

CONCLUSIONS

The removal of BC accompanying the removal of excessglycerol before freezing plus the destruction of WBCsduring freezing and deglycerolization reduces residualWBC counts to nearly 107 WBCs per unit, although re-maining above the current US standard of 5 � 106 WBCsper unit. This standard is achieved with removal of 20 mLof BC at the time of component separation with the useof a bag with a tapered top near the outlet port. How-ever, freezing and deglycerolization without priorWBC reduction does reduce WBC counts to nearly 107,and the deglycerolization also removes soluble anti-gens, toxins, and cytokines. Such a product should besatisfactory for all but a very few recipients9,16 for whomfiltration either before or after washing would be appro-priate.

ACKNOWLEDGMENTS

We thank Ralph Syring and Marne Hornblower for their assis-

tance in this project and acknowledge K. Banaudha, PhD, for

his help with apoptotic assays. We are grateful to the NIH and

NNMC blood banks for providing us with blood.

REFERENCES

1. Bordin JO, Heddle NM. Biologic effects of leukocytes

present in transfused cellular blood products. Blood 1994;

84:1703-21.

2. Frabetti F, Musiani D, Marini M, et al. White cell apopto-

sis in packed red cells. Transfusion 1998;38:1082-9.

3. Dzik WH. Effects on recipients of exposure to allogeneic

donor leukocytes. J Clin Apheresis 1994;9:135-8.

4. Heddle NM, Klama L, Singer J, et al. The role of the

plasma from platelet concentrates in transfusion reac-

tions. N Engl J Med 1994;331:625-8.

5. Vamvakas EC, Blajchman MA. Prestorage versus poststor-

age white cell reduction for the prevention of the deleteri-

ous immunomodulatory effects of allogeneic blood trans-

fusion. Transfus Med Rev 2000;14:23-33.

6. Novotny VM, van Doorn R, Witvliet MD, Claas FH, Brand

A. Occurrence of allogeneic HLA and non-HLA antibodies

after transfusion of prestorage filtered platelets and red

blood cells: a prospective study. Blood 1995;85:1736-41.

7. Goldman M, Delage G. The role of leukodepletion in the

control of transfusion-transmitted disease. Transfus Med

Rev 1995;9:9-19.

8. Heddle NM. Pathophysiology of febrile nonhemolytic

transfusion reactions. Curr Opin Hematol 1999;6:420-6.

9. Heddle N, Kelton J. Febrile non-hemolytic transfusion

reactions. In: Popovsky MA, ed. Transfusion reactions #2.

Bethesda, MD: AABB Press, 2001:45-82.

10. Uhlmann EJ, Isgriggs E, Wallhermfechtel M, Goodnough

LT. Prestorage universal WBC reduction of RBC units

does not affect the incidence of transfusion reactions.

Transfusion 2001;41:997-1000.

11. Farrugia A, Shea N, Knowles S, et al. Cryopreservation of

red blood cells: effect of freezing on red cell quality and

residual lymphocyte immunogenicity. J Clin Pathol 1993;

46:742-5.

12. Meryman HT, Hornblower M. The preparation of red

cells depleted of leukocytes. Rev Eval Transfusion 1986;26:

101-6.

13. Kurtz SR, Valeri DA, Melaragno AJ. Leukocyte-poor red

blood cells prepared by the addition and removal of glyc-

erol from red blood cell concentrates stores at 4 C. Trans-

fusion 1981;21:435-42.

14. Meryman HT, Bross J, Lebovitz R. The preparation of leu-

kocyte-poor red blood cells: a comparative study. Trans-

fusion 1980;20:285-92.

15. Valeri CR, Ragno G, Pivacek LE, et al. A multicenter study

of in vitro and in vivo values in human RBCs frozen with

40-percent (wt/vol) glycerol and stored after deglyceroli-

zation for 15 days at 4 degrees C in AS-3: assessment of

RBC processing in the ACP 215. Transfusion 2001;41:933-9.

16. AuBuchon JP, Elfath MD, Popovsky MA, et al. Evaluation

of a new prestorage leukoreduction filter for red blood

cell units. Vox Sang 1997;72:101-6.

ARNAUD AND MERYMAN


17. Miller JP, Mintz PD. The use of leukocyte-reduced blood

components. Hematol Oncol Clin North Am 1995;9:69-

90.

18. Pietersz RN, van der Meer PF, Seghatchian MJ. Update on

leukocyte depletion of blood components by filtration.

Transfus Sci 1998;19:321-8.

19. Farrugia A, Tan Y, Romeo A, et al. Relative efficiency of

leucocyte removal procedures for the production of leu-

cocyte-poor red cell concentrates assessed by flow cytom-

etry. Vox Sang 1994;66:153-60.

20. Tom Sahines, Mission Medical, Inc. M1000 device techni-

cal manual. (personal communication).

21. Meryman HT, Hornblower M. A simplified procedure for

deglycerolizing red blood cells frozen in a high glycerol

concentration. Transfusion 1977;17:438-42.

22. Valeri CR. Simplification of the method for adding and

removing glycerol during freeze-perservation of human

red blood cells with the high or low glycerol methods.

Biochemical modification prior to freezing. Transfusion

1975;15:195-218.

23. Clinical laboratory medicine. McClathey KD, ed. Balti-

more: Lippincott, Williams Wilkins, 2001:1378.

24. Weisbach V, Wanke C, Zingsem J, Zimmermann R, Eck-

stein R. Cytokine generation in whole blood, leukocyte-

depleted and temporarily warmed red blood cell concen-

trates. Vox Sang 1999;76:100-6.

25. Martelli AM, Tazzari PL, Bortul R, et al. Nuclear matrix

protein is released from apoptotic white cells during cold

(1-6 degrees C) storage of concentrated red cell units and

might induce antibody response in multiply transfused

patients. Transfusion 2000;40:169-77.

26. Biffl WL, Moore EE, Offner PJ, et al. Plasma from aged

stored red blood cells delays neutrophil apoptosis and

primes for cytotoxicity: abrogation by poststorage wash-

ing but not prestorage leukoreduction. J Trauma 2001;50:

426-31.



wbc reduction in cryopreserved rbc units

Documents