- Email: [email protected]
Filtration induces changes in activity states and leucocyte populations
Transfusion and Apheresis Science 28 (2003) 319–327 www.elsevier.com/locate/transci
WhatÕs happening?
Filtration induces changes in activity states and leucocyte populations H. Trindade a
a,b
, H Carvalho b, G. Sousa c, J.A. Machado Caetano b, J. Seghatchian d,*
Centro de Histocompatibilidade do Sul, Campo dos M artires da P atria, 130, 1169-056 Lisboa, Portugal b Servicßo Imunologia F.C.M.L., Campo dos M artires da P atria, 130, 1169-056 Lisboa, Portugal c Centro Regional de Sangue de Lisboa, 1PS, Av. do Brasil, 53-Pav.17, 1749-005 Lisboa, Portugal d Blood Component Technology Consultancy, 50 Primrose Hill Road, London NW3 3AA, UK
Abstract The distribution of leucocyte subpopulations in platelet concentrates (PC) derived from pre-storage filtered plateletrich plasma (PRP), the cell suspension obtained by reverse filter washing and the post-filtered PC, were monitored by immunophenotyping analysis using CD3, CD20 and CD33. Leucocyte activation analysis with the CD11b marker revealed that this molecule is up regulated in neutrophils taken from the filter. This, together with the loss of cell viability during the enrichment process, suggests that contact with the filter matrix and processing and storage of samples containing leucocytes may lead to activation and loss of leucocyte viability. These changes were found to be more pronounced in less stable myeloid cells and account for the differences reported among various authors which in some cases related to operational conditions such as the enrichment process used and the length of time between filtration and analysis of samples. Finally, statistical analysis of the results obtained by immunophenotypic studies indicate that postfilter samples (S) contain significantly higher numbers of CD33+ myeloid cells when compared to (PF) the pre-filter samples (65.03% 12.6 and 24.56% 14.73, p < 0:0000), with a decrease in T cells (50.72% 14.80 in PF and 24.05 9.48 in the cell suspension (S), p < 0:0007) and B cells (14.96 9.31 in PF and 9.9 5.22 in S, p < 0:201). A new strategy for assessing the influence of the filtration process on residual leucocyte activation and viability is described. This has direct relevance to collection, processing, storage and quality monitoring of PC. Ó 2003 Elsevier Science Ltd. All rights reserved.
1. Introduction The presence of leucocytes in blood component units is responsible for the majority of adverse
*
Corresponding author. Tel./fax: +44-0207-722-9596. E-mail address: [email protected] (J. Seghatchian).
transfusion reactions. These include HLA alloimmunization directly connected to non-hemolytic febrile reactions and platelet refractoriness, and transmission of microorganisms, among others. Therefore, the removal of leucocytes in certain clinical situations according to guidelines, by selective leucoreduction––is part of the approach to providing the patient with the best blood product. This takes into account the scientific evidence provided by a great number of published studies.
1473-0502/03/$ - see front matter Ó 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S1473-0502(03)00052-1
320
H. Trindade et al. / Transfusion and Apheresis Science 28 (2003) 319–327
Bearing in mind that both the length of storage before the separation of platelets and the relatively high residual remaining in components produced by some of the currently available technologies can influence the development and release of substances such as cytokines and neutrophil enzymes, it has been recommended that universal leucoreduction using the new generation of 4 Log reduction filters or processes prior to storage is the best way to prevent or minimize adverse transfusion reactions and complications [1,2]. Besides these facts and difficulties, the scientific data on the abnormal prion protein as the etiologic agent of BSE, which is the same agent as for the human disease, which is called variant Creutzfeldt– Jacob disease [3,4] and that it circulated in blood B cells before neuroinvasion [5], provided a theoretical risk of transmission of the new variant Creutzfeldt–Jacob disease through blood transfusion. These provided additional reasons for studying which leucocyte populations are reduced and how or which eventual factors may influence greater or lesser filtration effectiveness for certain populations. Conditions were created for several countries to adopt pre-storage leucodepletion procedures, for each and every cellular blood component [6], i.e., universal leucodepletion. In our situation platelet component therapy is mainly based on whole blood collection either through PRP or BC procedures. For this study we focused on the PRP procedure using medical devices with fourth generation PALL filters, which are recognized to significantly reduce the number of leucocytes and the incidence of transfusion-related adverse reactions and complications. Despite its capacity of leucodepletion, the effectiveness has been discussed in the regard to the filtration of different leucocyte populations. Some reports hint at greater effectiveness in neutrophil removal [7,8] and others at lymphocyte removal [9–11] or both without apparent consensus between the different studies. This preliminary report focused on providing additional evidence on the possible loss of leucocyte viability during the enrichment process as well as granulocyte activation subsequent to contact with the filter and changes in leucocyte subpopulations upon conventional filtration processes.
2. Materials and methods We studied 33 random donor platelet concentrates (PC) that had been leucoreduced pre-storage, by using a fourth generation PALL filter. From these, we obtained 19 pre-filtered plasma samples (PF), and 19 cell samples from the filter (F) for immunophenotyping studies. Of these, only eight contained leucocyte numbers that allowed post-filtering cell studies (S). The filter was reversely washed twice with 2.5 ml of PBS, cells were centrifuged 100 at 500g at 20 °C, and cells were then re-suspended in 2 ml of PBS. Pre-filter cells were obtained from platelet-rich plasma (PRP) that was in the medical device before the filter. These cells were re-suspended with 5 ml of PBS and centrifuged 100 at 500g at 20 °C while re-suspended in 1 ml of PBS. Residual white cells were counted using the Leucocount method (Becton Dickinsonâ ) [12]. Leucocyte purification from the leucodepleted concentrate was made through centrifugation at 700g for 10 min. The supernatant was then aspired, leaving 5 ml of plasma to re-suspend the cells. The cellular suspension was then centrifuged over 6 ml of Ficoll density gradient (D ¼ 1.056– 1.058 g/ml at 800g during 20 min). Cells were washed twice with 10 ml of PBS 5% albumin, and, finally, cells were re-suspended in 500 ll of PBS 5% albumin. To characterize residual cells, we used a direct immunofluorescent technique with the following monoclonal markers obtained from Beckton Dikinsonâ : CD33 (APC) for myeloid cells, CD3 (PE) for T cells, CD20 (PerCP) for B cells and pan leucocyte marker CD45 (FITC). A CD11b (PE) marker was only used for cell activation studies of CD33+ cells in F and PF cells. Briefly, 200 ll of the cell suspension were incubated for 20 min at room temperature with the above-listed combinations of monoclonal antibodies. After, 2 ml of FACs lysing solution were added and another 10 min incubation was performed in the dark. After this period, cells were washed once in PBS and re-suspended in 200 ll PBS 1% paraformaldehyde and stored at 4 °C until being analyzed in the flow cytometer.
H. Trindade et al. / Transfusion and Apheresis Science 28 (2003) 319–327
A minimum of 5000 events per tube were acquired on a FACSCalibur flow cytometer (Beckton Dickinsonâ ). Forward Scatter (FSC) and Side Scatter (SSC) were used to gate on viable lymphocytes, monocytes and granulocytes. The gated populations were analyzed by four-color immunofluorescence with Paint-a-Gate software (Beckton Dickinsonâ ).
321
number of cells to enable leucocyte characterization studies (Table 1). Therefore, that assessment was only possible in samples that showed a high number of residual cells, as well as a high recovery after purification in a Ficoll density gradient. This centrifugation in Ficoll density gradient, intended to eliminate platelets led to an increased loss of leucocytes in all studies. An example of cell loss and the percent recovery in four samples, is shown in Table 2.
3. Statistical methods Results are expressed as the mean standard deviation (M SD) of percentage of cells expressing a given phenotype. For non-parametric variables, the Mann–Whitney U test was used to investigate the potential existence of statistically significant differences between two or more groups. Statistical significance was defined as p < 0:05.
4. Results 4.1. The levels of the residual leucocytes in platelet concentrates and possible loss of cell viability upon enrichment process In the study of 33 units of PC obtained from random donor PRP we verified that the total number of cells varied between a minimum of 0 and a maximum of 1.26 106 , indicating considerable variation in the level of residual leucocytes from sample to sample, often with an insufficient
4.2. Possible activation of granulocytes upon contact with the filter material It is unclear to what extend the loss of leucocytes is related to changes in activation and viability of leucocytes. This was investigated on samples derived from the filter where cells come in close contact with the filter fibers, using CD11b expression of granulocytes as a measure of activation. Our results of a comparative study of CD11b expression in pre-filtration granulocytes in PC and in granulocytes obtained after filter reversed washing (less than 8 h after platelet in-line filtration) expressed as mean channel results, indicate that there is clear activation of the granulocytes (Fig. 1), with a mean channel rise from 788.1 to 1711.9 (n ¼ 19; p < 0:0000). 4.3. The distribution of leucocyte subpopulations in samples containing high and low levels of leucocytes As the cell populations in the pre-filter PRP may vary from sample to sample, we studied 10
Table 1 Sequential counting of 33 deleucocyted platelet concentrates with the purpose of showing variability verified cs (ll)
Total cs
cs (ll)
Total cs
cs (ll)
Total cs
0.055 0.327 1.58 0 0.327 0.218 21.746 0.109 0.545 19.384 0.273
3080 22890 82160 0 17985 11990 1261239 5995 29975 1160850 15015
0.109 0.436 0.327 1.119 0.218 1.615 0.143 0.048 0.048 0.238 2.23
5559 23544 16350 74338 11554 90440 8550 2565 2784 12376 115960
0.1 0.14 0.47 0.14 0.24 0.09 0.29 0.14 0.285 0.096 0.238
6600 8820 28200 8400 13680 3330 16200 8540 13680 4465 12138
322
H. Trindade et al. / Transfusion and Apheresis Science 28 (2003) 319–327
Table 2 Four examples of leucocount counting in platelet concentrates before and after density gradient centrifugation and percentage recovery calculated according to previous description [1] (% recovery ¼ post-enrichment WBC final volume 100/pre-enriched WBC processed volume) Four samples example
Cells before gradient centrifugation
Cells after gradient centrifugation
% Recovery
1 2 3 4
82160 28200 17985 1160850
11250 2600 5341 105200
13.7 9.2 29.7 9.1
the variability of the pre-filter PRP cell numbers among samples, and the variability of leucocyte lymphoid and myeloid populations, we studied the percentage of the main leucocyte populations in a group of 10 samples with the highest number of cells, and another group of 10 samples with the lowest number of cells, in order to determine whether the higher or lower cell number depended on a higher or lower yield of a certain leucocyte population (Table 4). The results show that in both groups there is a similar percentage of myeloid cells, and although the higher cell number group had a higher percentage of T lymphocytes and a lower percentage of B lymphocytes, there was no statistical significance. Considering that the cells obtained from reverse filter washing could represent a larger capacity for retaining certain leucocyte populations, we studied 19 samples of cell suspensions obtained by filter reverse wash. These were compared with corresponding samples obtained from pre-filter PRP (Table 5). Our results show a higher percentage of
Fig. 1. Mean channel (MC) amplification of CD11b granulocyte expression, in cells obtained from pre-filtration plasma (PF), and from filter (F) washing.
consecutive samples of pre-filter platelet-rich plasma. We found high variability with the T lymphocyte percentage varying from 19% to 54%, myeloid cells from 7% to 61% and B lymphocytes varying from 9.9% to 30% (Table 3). Considering
Table 3 Phenotype study of leucocyte populations in platelet-rich plasma of 10 successive samples % Cells
1
2
3
4
5
6
7
8
9
10
CD3+ CD33+ CD20+
31 25 15
54 8.1 15
23 39 13
35 28 9.9
45 9.5 24
36 35 15
19 61 11
32 38 14
40 7 30
48 13 21
Table 4 Main leucocyte populations in 10 samples with highest cell count and 10 with lowest cell count (statistical significance p < 0:05) Highest cell count n ¼ 10 (mean 351531.7) Lowest cell count n ¼ 10 (mean 6342.8) p value
CD3+
CD33+
CD20+
52.6 11.4 40.0 9.3 0.0525
20.6 8.4 20.1 12.6 0.9391
14.9 8.8 18.9 9.1 0.3763
H. Trindade et al. / Transfusion and Apheresis Science 28 (2003) 319–327
323
Table 5 Comparative study of leucocyte populations in two groups of samples obtained from reverse filter washing and from platelet-rich plasma before filtration Pre-filter (PF) n ¼ 19 Filter reverse wash (F) n ¼ 19 p-values
CD3
CD33
CD20
42.1 12.9 43.7 10.8 0.2447
23.9 14.1 20.3 11.3 0.0926
16.5 7.5 19.8 6.1 0.0078
B cells in the cells obtained from the filter, with discrete predominance of myeloid cells in PRP. T lymphocytes show no difference. 4.4. Comparison of leucocyte subpopulation in pre/ post-filtered samples and the filter content As referred to previously, out of the 33 samples of PC, only eight had enough cells to allow a comparative analysis on the three sets of samples
from the same source (PF, F and S). As shown in Table 6, the results of this comparative analysis revealed that there is no significant difference in the percentage of myeloid and T cell markers (PFCD33 ¼ 24.56 14.73/F-CD33 ¼ 21.95 14.55, p < 0:7265 and PF-CD3 ¼ 50.72 14.80/F-CD3 ¼ 48.97 11.62, p < 0:7964). The B cells showed the highest percentage in samples F (21.86 10.88), when compared to PF (14.96 9.32, p < 0:194). It is clear that the percentage of myeloid cells is
Table 6 Summary of results of plasma studies before filtering (PF), in cells obtained by reverse filter washing (F) and in residual cells of platelet concentrate after filtering (S) CD20 (%)
CD3 (%)
CD33 (%)
Sample 1
PF F S
12.0 16.8 6.5
40.2 51.1 20.9
26.0 21.1 72.5
Sample 2
PF F S
9.1 17.7 1.1
69.1 61.6 6.5
14.5 9.8 92.2
Sample 3
PF F S
34.8 29.9 15
40.5 43.8 29.3
7.9 14.7 54.2
Sample 4
PF F S
5.7 7.9 7.9
49.1 57.9 37.8
25.3 13.6 53.7
Sample 5
PF F S
10.0 16.4 13.3
41.2 52.1 18.6
34.0 22.1 66.6
Sample 6
PF F S
22.0 40.1 7.1
43.1 34.4 31.6
18.5 24.8 58.1
Sample 7
PF F S
11.6 18.0 17.0
69.0 60.5 21.8
18.3 19.6 61.1
Sample 8
PF F S
12.1 31.8 11.3
63.9 50.5 25.9
22.3 15.1 61.9
324
H. Trindade et al. / Transfusion and Apheresis Science 28 (2003) 319–327
Table 7 Results of the study of leucocyte populations in two groups of samples obtained from pre-filter PRP and from post-filtration plasma of the same sample units PF% n ¼ 8 S% n ¼ 8 p-values
CD3
CD33
CD20
50.72 14.80 24.05 9.48 0.0007
24.56 14.73 65.03 12.6 0.0000
14.96 9.31 9.90 5.22 0.201
higher in both samples which come in contact with the filter with a decrease of both B and T cell percentages (Table 7). Leucocytes obtained from sample S, when compared to F show a lower percentage of T cells (S-CD3 ¼ 24.05 9.48/F-CD3 ¼ 48.97 11.62, p < 0:0003), and myeloid (F-CD33 ¼ 21.95 14.55 and S-CD33 ¼ 65.03 12.62, p < 0:0000) might be expected, they have a lower percentage of B cells (F-CD20 ¼ 21.86 10.88 and S-CD20 ¼ 9.90 5.22, p < 0:0141).
5. Discussion In our experience the final number of residual leucocytes varies sharply from sample to sample, and in most cases we did not have a sufficient number of cells to perform the study of postfiltration leucocyte populations. This agrees with information from others [8,11]. Even in samples that initially displayed a high number of leucocytes that might enable the performance of post-filtering studies, we verified the fact that after concentration in Ficoll density gradient, the number decreased sharply, with a low percent recovery. Moreover, the residual leucocytes often displayed losses of their Side and Forward Scatter characteristics in dot plot FS SS (Fig. 2) which we interpreted as an ongoing cellular apoptosis/necrosis process. This lead us to from the hypothesis that diminishing cell viability is the cause for the loss seen during these studies. Nevertheless we cannot exclude that the cell loss may be due to loss of leucocyte viability induced by trauma caused by contact with the filter fibers [13]. In that case, the timing of sample studies may play an important role in the determining results observed, as longer periods of time between collection, filtration and
studies may correspond to higher cell loss and this may be more pronounced in certain leucocyte populations. Sample study time varies greatly among different authors, ranging from days [7] to hours [11] as was our case, with all samples being studied up to 8 h after filtration. Besides filtration trauma, cell viability may be affected by other factors such as the time of processing of the whole blood unit and the development of ghost cells, and nucleus and cytoplasmatic materials resulting from dead cells scattered in the plasma as reported by others [14,15]. Cell activation is another process that may contribute both to cell viability changes and filtration efficiency. Some authors infer that HLA DR expression leads to cells being differently distributed along different filter layers [13] and this would lead to the activated cells having a different adherence ability. In our study, T cellular activation was studied with CD69 and CD25 markers (data not shown) in cells obtained through filter reverse washing and we found no significant alterations when compared to the expression of the same markers in pre-filter plasma cells. Our results with granulocyte CD11b expression, however, point to an up-modulation of this surface marker in cells obtained from filters. As we do not have sufficient cell numbers to perform the same study in cells after filtration, we can presume that this post-filter population is also activated by contact with the filter fibers. If so, this activation may induce degranulation and side scatter alterations as well as a low half-life when compared with prefilter cells, which is consistent with our results. Another issue which has arisen has to do with quantitative and qualitative variability in matters of lymphoid or myeloid cell predominance in the pre-filter PRP. This variability affects processing
H. Trindade et al. / Transfusion and Apheresis Science 28 (2003) 319–327
325
Fig. 2. Forward side scatter alterations in post-filtration cells identified by their phenotype characteristics.
from the blood unit through to the PRP stage, and may affect the results of this type of study. In fact, in a successive study of 10 samples we verified that there are sharp variations of the relative percentages of different leucocyte populations in the pre-filtration PRP, with some showing a larger predominance of lymphoid cells and others of myeloid cells. These differences might suggest that a higher percentage of myeloid cells in the pre-filtration PRP might correspond to a higher percentage of these cells in the post-filter PRP. We nevertheless verified that, in samples where we were able to perform joint phenotyping in pre- and post-filtration samples, the myeloid cells showed high percentages post filtration independently of their initial percentage (Table 6, samples 2 and 6). Therefore, our results point to no great influence of this variability factor in determining the final cell percentage as shown for B cells by others [10]. Cells obtained from filter washing might contribute to a better knowledge of which cells are more easily retained in the fibers, and these results might be compared with those from post-filtration. However, we consider it possible that the study of leucocyte subpopulations obtained through filter washing may be influenced by two different
situations, one having to do with the critical force needed to release the adhering cells, and the other to the higher or lower filtration efficiency for each of the populations. A higher percentage of B cells in the reverse filter washing cell suspension may suggest both that these cells are more efficiently filtered cells and that they may be released more easily. Both T CD3+ cells and CD33+ myeloid cells show no significant differences. If the greater efficiency for these populations proved to be true, one might expect that in the study performed on final samples (S) there should be a decrease of B cells in the post-filtration PRP when compared to the pre-filtration results. In fact, our studies clearly show that B cells do decrease after filtration as shown by others [10], while there is a greater percentage of myeloid CD33+ cells. T cells nevertheless show different behavior from B cells as, when comparing the pre-filter with the filter wash they do not show meaningful differences, whereas they show a significant decrease in the post-filter. As we are expressing results in percentages, the decrease of this population may follow the percentage increase of myeloid cells. Despite that, it seems we may point to the different behavior of T and B populations, in which both would be efficiently
326
H. Trindade et al. / Transfusion and Apheresis Science 28 (2003) 319–327
retained by the filter, but it would be more difficult to elude T cells by a reverse wash. Considering our results and observing each case (Table 6) we see that the variability in the presence of cells in the pre-filter does not seem to have a direct relationship to the efficiency of filtration, with the myeloid population being the one that appears in highest percentage in the post-filtration concentrate. This data agrees with what is described by other authors who found a fourfold (2.6 in our experience) neutrophil concentration postfiltration [11] and does not agree with what has been reported by others [7,8]. Nevertheless one must keep in mind that the time period that elapses between filtering and subsequent analysis may lead to different results and that myeloid cells would suffer a degenerative process.
6. Conclusions The assessment of leucocyte viability and activation states are crucially important to ensure the consistency in performance of the filtration process but the study of cell populations in leucoreduced PC is made difficult by the low number of residual cells. This low cell number seems to be due not only to filtration efficiency, but also to possible alterations of the cell viability with an ensuing loss during cell enrichment manipulation. Comparative analysis of pre- and post-filtration samples revealed that the T and B cells seem to be efficiently retained by the filter. Moreover, in our model, myeloid cells appeared in higher percentage after filtration. These data suggest that changes in leucocyte subpopulations and the activation state of granulocytes occur subsequent to cell enrichment or filtration processes. It remains to be established as to what degree these changes are due to the contact of the leucocytes with the filter fibers and loss of viability by trauma or the period of time that elapses between filtration and subsequent analysis which can grossly influence the overall outcome. Further work is in progress in order to clarify the relationship between time and the number of live and apoptotic cells that may be present in some pre- and post-filtration samples, as well as the phenotypic alterations of myeloid cells
during conventional collection/processing and storage.
References [1] International forum. Universal leucocyte-depletion of blood components: cell concentrates and plasma. Vox Sang 2001;81:56–77. [2] Dzik S, AuBuchon J, Jeffries L, Kleinman S, Manno C, Murphy M, et al. Leucocyte reduction of blood components: public policy and new technology. Transfus Med Rev 2000;14:34–52. [3] Bruce ME, Will RG, Ironside JW, McConnell I, Drummond D, Suttie A, et al. Transmissions to mice indicate that ‘‘new variant’’ CJD is caused by the BSE agent. Nature 1997;389:498–501. [4] Hill AF, Desbruslais M, Joiner S, Sidle KCL, Gowland I, Doey LJ, et al. The same prion strain causes vCJD and BSE. Nature 1997;389:448–50. [5] Klein MA, Frigg R, Flechsig E, Raeber AJ, Kalinke U, Bluethmann H, et al. A crucial role for B cells in neuroinvasive scrapie. Nature 1997;390:687–90. [6] Murphy MF. New variant Creutzfeldt–Jacob Disease (nvCJD): the risk of transmission by blood transfusion and the potential benefit of leucocyte reduction of blood components. Transfus Med Rev 1999;13:75–83. [7] Sowemimo-Coker SO, Kirn A, Tribble E, Brandwein HJ, Wenz B. White cell subsets in apheresis and filtered platelet concentrates. Transfusion 1998;38:650–7. [8] Seghatchian J, Krailadsiri P. Validation of different enrichment strategies for analysis of leucocyte sub-population: development and application of a new approach, based on leucofiltration. Transfus Apher Sci 2002;26:61– 72. [9] Rider JR, Want EJ, Winter MA, Turton JR, Pamphilon DH, Nobes P. Differential leucocyte subpopulation analysis of leucodepleted red cell products. Transf Med 2000;10:49–58. [10] Mascaretti L, Baggi L, Riva M, Proserpio P, Dassi M, Varallo F, et al. Lymphocyte subsets in inline packed red blood cell units: comparison between low and high spin procedures. Tranfus Apher Sci 2002;26:167–74. [11] Triulzi DJ, Meyer EM, Donnenberg AD. WBC subsets analysis of WBC-reduced platelet concentrates. Transfusion 2000;40:771–80. [12] Barclay GR, Walker B, Gibson J, McColl K, Turner ML. Quality assurance by a commercial flow cytometry method of leucodepletion of whole blood donations: initial application of universal testing and proposals for a batch-release sampling plan. Transf Med 2000;10:37–48. [13] Steneker I, Beiwenga J. Histologic and immunohistochemical studies on the preparation of white cell-poor red cell concentrates: the filtration process using three different polyester filters. Transfusion 1991;31:40–6.
H. Trindade et al. / Transfusion and Apheresis Science 28 (2003) 319–327 [14] Sivakumaran M, Saunders P, Hanmer D, Drury M, Seghatchian J. Leucocyte fragments in prestorage-leucodepleted blood products: a qualitative analysis. Transfus Sci 2000;22:81–2.
327
[15] Seghatchian J, Krailadsiri P, Beard M, Bukasa A, Bashir S. Studies on the characterization of the cause of leucoreduction failures with particular reference to extra gatal events. Transfus Apher Sci 2002;26:47–60.
WhatÕs happening?
Filtration induces changes in activity states and leucocyte populations H. Trindade a
a,b
, H Carvalho b, G. Sousa c, J.A. Machado Caetano b, J. Seghatchian d,*
Centro de Histocompatibilidade do Sul, Campo dos M artires da P atria, 130, 1169-056 Lisboa, Portugal b Servicßo Imunologia F.C.M.L., Campo dos M artires da P atria, 130, 1169-056 Lisboa, Portugal c Centro Regional de Sangue de Lisboa, 1PS, Av. do Brasil, 53-Pav.17, 1749-005 Lisboa, Portugal d Blood Component Technology Consultancy, 50 Primrose Hill Road, London NW3 3AA, UK
Abstract The distribution of leucocyte subpopulations in platelet concentrates (PC) derived from pre-storage filtered plateletrich plasma (PRP), the cell suspension obtained by reverse filter washing and the post-filtered PC, were monitored by immunophenotyping analysis using CD3, CD20 and CD33. Leucocyte activation analysis with the CD11b marker revealed that this molecule is up regulated in neutrophils taken from the filter. This, together with the loss of cell viability during the enrichment process, suggests that contact with the filter matrix and processing and storage of samples containing leucocytes may lead to activation and loss of leucocyte viability. These changes were found to be more pronounced in less stable myeloid cells and account for the differences reported among various authors which in some cases related to operational conditions such as the enrichment process used and the length of time between filtration and analysis of samples. Finally, statistical analysis of the results obtained by immunophenotypic studies indicate that postfilter samples (S) contain significantly higher numbers of CD33+ myeloid cells when compared to (PF) the pre-filter samples (65.03% 12.6 and 24.56% 14.73, p < 0:0000), with a decrease in T cells (50.72% 14.80 in PF and 24.05 9.48 in the cell suspension (S), p < 0:0007) and B cells (14.96 9.31 in PF and 9.9 5.22 in S, p < 0:201). A new strategy for assessing the influence of the filtration process on residual leucocyte activation and viability is described. This has direct relevance to collection, processing, storage and quality monitoring of PC. Ó 2003 Elsevier Science Ltd. All rights reserved.
1. Introduction The presence of leucocytes in blood component units is responsible for the majority of adverse
*
Corresponding author. Tel./fax: +44-0207-722-9596. E-mail address: [email protected] (J. Seghatchian).
transfusion reactions. These include HLA alloimmunization directly connected to non-hemolytic febrile reactions and platelet refractoriness, and transmission of microorganisms, among others. Therefore, the removal of leucocytes in certain clinical situations according to guidelines, by selective leucoreduction––is part of the approach to providing the patient with the best blood product. This takes into account the scientific evidence provided by a great number of published studies.
1473-0502/03/$ - see front matter Ó 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S1473-0502(03)00052-1
320
H. Trindade et al. / Transfusion and Apheresis Science 28 (2003) 319–327
Bearing in mind that both the length of storage before the separation of platelets and the relatively high residual remaining in components produced by some of the currently available technologies can influence the development and release of substances such as cytokines and neutrophil enzymes, it has been recommended that universal leucoreduction using the new generation of 4 Log reduction filters or processes prior to storage is the best way to prevent or minimize adverse transfusion reactions and complications [1,2]. Besides these facts and difficulties, the scientific data on the abnormal prion protein as the etiologic agent of BSE, which is the same agent as for the human disease, which is called variant Creutzfeldt– Jacob disease [3,4] and that it circulated in blood B cells before neuroinvasion [5], provided a theoretical risk of transmission of the new variant Creutzfeldt–Jacob disease through blood transfusion. These provided additional reasons for studying which leucocyte populations are reduced and how or which eventual factors may influence greater or lesser filtration effectiveness for certain populations. Conditions were created for several countries to adopt pre-storage leucodepletion procedures, for each and every cellular blood component [6], i.e., universal leucodepletion. In our situation platelet component therapy is mainly based on whole blood collection either through PRP or BC procedures. For this study we focused on the PRP procedure using medical devices with fourth generation PALL filters, which are recognized to significantly reduce the number of leucocytes and the incidence of transfusion-related adverse reactions and complications. Despite its capacity of leucodepletion, the effectiveness has been discussed in the regard to the filtration of different leucocyte populations. Some reports hint at greater effectiveness in neutrophil removal [7,8] and others at lymphocyte removal [9–11] or both without apparent consensus between the different studies. This preliminary report focused on providing additional evidence on the possible loss of leucocyte viability during the enrichment process as well as granulocyte activation subsequent to contact with the filter and changes in leucocyte subpopulations upon conventional filtration processes.
2. Materials and methods We studied 33 random donor platelet concentrates (PC) that had been leucoreduced pre-storage, by using a fourth generation PALL filter. From these, we obtained 19 pre-filtered plasma samples (PF), and 19 cell samples from the filter (F) for immunophenotyping studies. Of these, only eight contained leucocyte numbers that allowed post-filtering cell studies (S). The filter was reversely washed twice with 2.5 ml of PBS, cells were centrifuged 100 at 500g at 20 °C, and cells were then re-suspended in 2 ml of PBS. Pre-filter cells were obtained from platelet-rich plasma (PRP) that was in the medical device before the filter. These cells were re-suspended with 5 ml of PBS and centrifuged 100 at 500g at 20 °C while re-suspended in 1 ml of PBS. Residual white cells were counted using the Leucocount method (Becton Dickinsonâ ) [12]. Leucocyte purification from the leucodepleted concentrate was made through centrifugation at 700g for 10 min. The supernatant was then aspired, leaving 5 ml of plasma to re-suspend the cells. The cellular suspension was then centrifuged over 6 ml of Ficoll density gradient (D ¼ 1.056– 1.058 g/ml at 800g during 20 min). Cells were washed twice with 10 ml of PBS 5% albumin, and, finally, cells were re-suspended in 500 ll of PBS 5% albumin. To characterize residual cells, we used a direct immunofluorescent technique with the following monoclonal markers obtained from Beckton Dikinsonâ : CD33 (APC) for myeloid cells, CD3 (PE) for T cells, CD20 (PerCP) for B cells and pan leucocyte marker CD45 (FITC). A CD11b (PE) marker was only used for cell activation studies of CD33+ cells in F and PF cells. Briefly, 200 ll of the cell suspension were incubated for 20 min at room temperature with the above-listed combinations of monoclonal antibodies. After, 2 ml of FACs lysing solution were added and another 10 min incubation was performed in the dark. After this period, cells were washed once in PBS and re-suspended in 200 ll PBS 1% paraformaldehyde and stored at 4 °C until being analyzed in the flow cytometer.
H. Trindade et al. / Transfusion and Apheresis Science 28 (2003) 319–327
A minimum of 5000 events per tube were acquired on a FACSCalibur flow cytometer (Beckton Dickinsonâ ). Forward Scatter (FSC) and Side Scatter (SSC) were used to gate on viable lymphocytes, monocytes and granulocytes. The gated populations were analyzed by four-color immunofluorescence with Paint-a-Gate software (Beckton Dickinsonâ ).
321
number of cells to enable leucocyte characterization studies (Table 1). Therefore, that assessment was only possible in samples that showed a high number of residual cells, as well as a high recovery after purification in a Ficoll density gradient. This centrifugation in Ficoll density gradient, intended to eliminate platelets led to an increased loss of leucocytes in all studies. An example of cell loss and the percent recovery in four samples, is shown in Table 2.
3. Statistical methods Results are expressed as the mean standard deviation (M SD) of percentage of cells expressing a given phenotype. For non-parametric variables, the Mann–Whitney U test was used to investigate the potential existence of statistically significant differences between two or more groups. Statistical significance was defined as p < 0:05.
4. Results 4.1. The levels of the residual leucocytes in platelet concentrates and possible loss of cell viability upon enrichment process In the study of 33 units of PC obtained from random donor PRP we verified that the total number of cells varied between a minimum of 0 and a maximum of 1.26 106 , indicating considerable variation in the level of residual leucocytes from sample to sample, often with an insufficient
4.2. Possible activation of granulocytes upon contact with the filter material It is unclear to what extend the loss of leucocytes is related to changes in activation and viability of leucocytes. This was investigated on samples derived from the filter where cells come in close contact with the filter fibers, using CD11b expression of granulocytes as a measure of activation. Our results of a comparative study of CD11b expression in pre-filtration granulocytes in PC and in granulocytes obtained after filter reversed washing (less than 8 h after platelet in-line filtration) expressed as mean channel results, indicate that there is clear activation of the granulocytes (Fig. 1), with a mean channel rise from 788.1 to 1711.9 (n ¼ 19; p < 0:0000). 4.3. The distribution of leucocyte subpopulations in samples containing high and low levels of leucocytes As the cell populations in the pre-filter PRP may vary from sample to sample, we studied 10
Table 1 Sequential counting of 33 deleucocyted platelet concentrates with the purpose of showing variability verified cs (ll)
Total cs
cs (ll)
Total cs
cs (ll)
Total cs
0.055 0.327 1.58 0 0.327 0.218 21.746 0.109 0.545 19.384 0.273
3080 22890 82160 0 17985 11990 1261239 5995 29975 1160850 15015
0.109 0.436 0.327 1.119 0.218 1.615 0.143 0.048 0.048 0.238 2.23
5559 23544 16350 74338 11554 90440 8550 2565 2784 12376 115960
0.1 0.14 0.47 0.14 0.24 0.09 0.29 0.14 0.285 0.096 0.238
6600 8820 28200 8400 13680 3330 16200 8540 13680 4465 12138
322
H. Trindade et al. / Transfusion and Apheresis Science 28 (2003) 319–327
Table 2 Four examples of leucocount counting in platelet concentrates before and after density gradient centrifugation and percentage recovery calculated according to previous description [1] (% recovery ¼ post-enrichment WBC final volume 100/pre-enriched WBC processed volume) Four samples example
Cells before gradient centrifugation
Cells after gradient centrifugation
% Recovery
1 2 3 4
82160 28200 17985 1160850
11250 2600 5341 105200
13.7 9.2 29.7 9.1
the variability of the pre-filter PRP cell numbers among samples, and the variability of leucocyte lymphoid and myeloid populations, we studied the percentage of the main leucocyte populations in a group of 10 samples with the highest number of cells, and another group of 10 samples with the lowest number of cells, in order to determine whether the higher or lower cell number depended on a higher or lower yield of a certain leucocyte population (Table 4). The results show that in both groups there is a similar percentage of myeloid cells, and although the higher cell number group had a higher percentage of T lymphocytes and a lower percentage of B lymphocytes, there was no statistical significance. Considering that the cells obtained from reverse filter washing could represent a larger capacity for retaining certain leucocyte populations, we studied 19 samples of cell suspensions obtained by filter reverse wash. These were compared with corresponding samples obtained from pre-filter PRP (Table 5). Our results show a higher percentage of
Fig. 1. Mean channel (MC) amplification of CD11b granulocyte expression, in cells obtained from pre-filtration plasma (PF), and from filter (F) washing.
consecutive samples of pre-filter platelet-rich plasma. We found high variability with the T lymphocyte percentage varying from 19% to 54%, myeloid cells from 7% to 61% and B lymphocytes varying from 9.9% to 30% (Table 3). Considering
Table 3 Phenotype study of leucocyte populations in platelet-rich plasma of 10 successive samples % Cells
1
2
3
4
5
6
7
8
9
10
CD3+ CD33+ CD20+
31 25 15
54 8.1 15
23 39 13
35 28 9.9
45 9.5 24
36 35 15
19 61 11
32 38 14
40 7 30
48 13 21
Table 4 Main leucocyte populations in 10 samples with highest cell count and 10 with lowest cell count (statistical significance p < 0:05) Highest cell count n ¼ 10 (mean 351531.7) Lowest cell count n ¼ 10 (mean 6342.8) p value
CD3+
CD33+
CD20+
52.6 11.4 40.0 9.3 0.0525
20.6 8.4 20.1 12.6 0.9391
14.9 8.8 18.9 9.1 0.3763
H. Trindade et al. / Transfusion and Apheresis Science 28 (2003) 319–327
323
Table 5 Comparative study of leucocyte populations in two groups of samples obtained from reverse filter washing and from platelet-rich plasma before filtration Pre-filter (PF) n ¼ 19 Filter reverse wash (F) n ¼ 19 p-values
CD3
CD33
CD20
42.1 12.9 43.7 10.8 0.2447
23.9 14.1 20.3 11.3 0.0926
16.5 7.5 19.8 6.1 0.0078
B cells in the cells obtained from the filter, with discrete predominance of myeloid cells in PRP. T lymphocytes show no difference. 4.4. Comparison of leucocyte subpopulation in pre/ post-filtered samples and the filter content As referred to previously, out of the 33 samples of PC, only eight had enough cells to allow a comparative analysis on the three sets of samples
from the same source (PF, F and S). As shown in Table 6, the results of this comparative analysis revealed that there is no significant difference in the percentage of myeloid and T cell markers (PFCD33 ¼ 24.56 14.73/F-CD33 ¼ 21.95 14.55, p < 0:7265 and PF-CD3 ¼ 50.72 14.80/F-CD3 ¼ 48.97 11.62, p < 0:7964). The B cells showed the highest percentage in samples F (21.86 10.88), when compared to PF (14.96 9.32, p < 0:194). It is clear that the percentage of myeloid cells is
Table 6 Summary of results of plasma studies before filtering (PF), in cells obtained by reverse filter washing (F) and in residual cells of platelet concentrate after filtering (S) CD20 (%)
CD3 (%)
CD33 (%)
Sample 1
PF F S
12.0 16.8 6.5
40.2 51.1 20.9
26.0 21.1 72.5
Sample 2
PF F S
9.1 17.7 1.1
69.1 61.6 6.5
14.5 9.8 92.2
Sample 3
PF F S
34.8 29.9 15
40.5 43.8 29.3
7.9 14.7 54.2
Sample 4
PF F S
5.7 7.9 7.9
49.1 57.9 37.8
25.3 13.6 53.7
Sample 5
PF F S
10.0 16.4 13.3
41.2 52.1 18.6
34.0 22.1 66.6
Sample 6
PF F S
22.0 40.1 7.1
43.1 34.4 31.6
18.5 24.8 58.1
Sample 7
PF F S
11.6 18.0 17.0
69.0 60.5 21.8
18.3 19.6 61.1
Sample 8
PF F S
12.1 31.8 11.3
63.9 50.5 25.9
22.3 15.1 61.9
324
H. Trindade et al. / Transfusion and Apheresis Science 28 (2003) 319–327
Table 7 Results of the study of leucocyte populations in two groups of samples obtained from pre-filter PRP and from post-filtration plasma of the same sample units PF% n ¼ 8 S% n ¼ 8 p-values
CD3
CD33
CD20
50.72 14.80 24.05 9.48 0.0007
24.56 14.73 65.03 12.6 0.0000
14.96 9.31 9.90 5.22 0.201
higher in both samples which come in contact with the filter with a decrease of both B and T cell percentages (Table 7). Leucocytes obtained from sample S, when compared to F show a lower percentage of T cells (S-CD3 ¼ 24.05 9.48/F-CD3 ¼ 48.97 11.62, p < 0:0003), and myeloid (F-CD33 ¼ 21.95 14.55 and S-CD33 ¼ 65.03 12.62, p < 0:0000) might be expected, they have a lower percentage of B cells (F-CD20 ¼ 21.86 10.88 and S-CD20 ¼ 9.90 5.22, p < 0:0141).
5. Discussion In our experience the final number of residual leucocytes varies sharply from sample to sample, and in most cases we did not have a sufficient number of cells to perform the study of postfiltration leucocyte populations. This agrees with information from others [8,11]. Even in samples that initially displayed a high number of leucocytes that might enable the performance of post-filtering studies, we verified the fact that after concentration in Ficoll density gradient, the number decreased sharply, with a low percent recovery. Moreover, the residual leucocytes often displayed losses of their Side and Forward Scatter characteristics in dot plot FS SS (Fig. 2) which we interpreted as an ongoing cellular apoptosis/necrosis process. This lead us to from the hypothesis that diminishing cell viability is the cause for the loss seen during these studies. Nevertheless we cannot exclude that the cell loss may be due to loss of leucocyte viability induced by trauma caused by contact with the filter fibers [13]. In that case, the timing of sample studies may play an important role in the determining results observed, as longer periods of time between collection, filtration and
studies may correspond to higher cell loss and this may be more pronounced in certain leucocyte populations. Sample study time varies greatly among different authors, ranging from days [7] to hours [11] as was our case, with all samples being studied up to 8 h after filtration. Besides filtration trauma, cell viability may be affected by other factors such as the time of processing of the whole blood unit and the development of ghost cells, and nucleus and cytoplasmatic materials resulting from dead cells scattered in the plasma as reported by others [14,15]. Cell activation is another process that may contribute both to cell viability changes and filtration efficiency. Some authors infer that HLA DR expression leads to cells being differently distributed along different filter layers [13] and this would lead to the activated cells having a different adherence ability. In our study, T cellular activation was studied with CD69 and CD25 markers (data not shown) in cells obtained through filter reverse washing and we found no significant alterations when compared to the expression of the same markers in pre-filter plasma cells. Our results with granulocyte CD11b expression, however, point to an up-modulation of this surface marker in cells obtained from filters. As we do not have sufficient cell numbers to perform the same study in cells after filtration, we can presume that this post-filter population is also activated by contact with the filter fibers. If so, this activation may induce degranulation and side scatter alterations as well as a low half-life when compared with prefilter cells, which is consistent with our results. Another issue which has arisen has to do with quantitative and qualitative variability in matters of lymphoid or myeloid cell predominance in the pre-filter PRP. This variability affects processing
H. Trindade et al. / Transfusion and Apheresis Science 28 (2003) 319–327
325
Fig. 2. Forward side scatter alterations in post-filtration cells identified by their phenotype characteristics.
from the blood unit through to the PRP stage, and may affect the results of this type of study. In fact, in a successive study of 10 samples we verified that there are sharp variations of the relative percentages of different leucocyte populations in the pre-filtration PRP, with some showing a larger predominance of lymphoid cells and others of myeloid cells. These differences might suggest that a higher percentage of myeloid cells in the pre-filtration PRP might correspond to a higher percentage of these cells in the post-filter PRP. We nevertheless verified that, in samples where we were able to perform joint phenotyping in pre- and post-filtration samples, the myeloid cells showed high percentages post filtration independently of their initial percentage (Table 6, samples 2 and 6). Therefore, our results point to no great influence of this variability factor in determining the final cell percentage as shown for B cells by others [10]. Cells obtained from filter washing might contribute to a better knowledge of which cells are more easily retained in the fibers, and these results might be compared with those from post-filtration. However, we consider it possible that the study of leucocyte subpopulations obtained through filter washing may be influenced by two different
situations, one having to do with the critical force needed to release the adhering cells, and the other to the higher or lower filtration efficiency for each of the populations. A higher percentage of B cells in the reverse filter washing cell suspension may suggest both that these cells are more efficiently filtered cells and that they may be released more easily. Both T CD3+ cells and CD33+ myeloid cells show no significant differences. If the greater efficiency for these populations proved to be true, one might expect that in the study performed on final samples (S) there should be a decrease of B cells in the post-filtration PRP when compared to the pre-filtration results. In fact, our studies clearly show that B cells do decrease after filtration as shown by others [10], while there is a greater percentage of myeloid CD33+ cells. T cells nevertheless show different behavior from B cells as, when comparing the pre-filter with the filter wash they do not show meaningful differences, whereas they show a significant decrease in the post-filter. As we are expressing results in percentages, the decrease of this population may follow the percentage increase of myeloid cells. Despite that, it seems we may point to the different behavior of T and B populations, in which both would be efficiently
326
H. Trindade et al. / Transfusion and Apheresis Science 28 (2003) 319–327
retained by the filter, but it would be more difficult to elude T cells by a reverse wash. Considering our results and observing each case (Table 6) we see that the variability in the presence of cells in the pre-filter does not seem to have a direct relationship to the efficiency of filtration, with the myeloid population being the one that appears in highest percentage in the post-filtration concentrate. This data agrees with what is described by other authors who found a fourfold (2.6 in our experience) neutrophil concentration postfiltration [11] and does not agree with what has been reported by others [7,8]. Nevertheless one must keep in mind that the time period that elapses between filtering and subsequent analysis may lead to different results and that myeloid cells would suffer a degenerative process.
6. Conclusions The assessment of leucocyte viability and activation states are crucially important to ensure the consistency in performance of the filtration process but the study of cell populations in leucoreduced PC is made difficult by the low number of residual cells. This low cell number seems to be due not only to filtration efficiency, but also to possible alterations of the cell viability with an ensuing loss during cell enrichment manipulation. Comparative analysis of pre- and post-filtration samples revealed that the T and B cells seem to be efficiently retained by the filter. Moreover, in our model, myeloid cells appeared in higher percentage after filtration. These data suggest that changes in leucocyte subpopulations and the activation state of granulocytes occur subsequent to cell enrichment or filtration processes. It remains to be established as to what degree these changes are due to the contact of the leucocytes with the filter fibers and loss of viability by trauma or the period of time that elapses between filtration and subsequent analysis which can grossly influence the overall outcome. Further work is in progress in order to clarify the relationship between time and the number of live and apoptotic cells that may be present in some pre- and post-filtration samples, as well as the phenotypic alterations of myeloid cells
during conventional collection/processing and storage.
References [1] International forum. Universal leucocyte-depletion of blood components: cell concentrates and plasma. Vox Sang 2001;81:56–77. [2] Dzik S, AuBuchon J, Jeffries L, Kleinman S, Manno C, Murphy M, et al. Leucocyte reduction of blood components: public policy and new technology. Transfus Med Rev 2000;14:34–52. [3] Bruce ME, Will RG, Ironside JW, McConnell I, Drummond D, Suttie A, et al. Transmissions to mice indicate that ‘‘new variant’’ CJD is caused by the BSE agent. Nature 1997;389:498–501. [4] Hill AF, Desbruslais M, Joiner S, Sidle KCL, Gowland I, Doey LJ, et al. The same prion strain causes vCJD and BSE. Nature 1997;389:448–50. [5] Klein MA, Frigg R, Flechsig E, Raeber AJ, Kalinke U, Bluethmann H, et al. A crucial role for B cells in neuroinvasive scrapie. Nature 1997;390:687–90. [6] Murphy MF. New variant Creutzfeldt–Jacob Disease (nvCJD): the risk of transmission by blood transfusion and the potential benefit of leucocyte reduction of blood components. Transfus Med Rev 1999;13:75–83. [7] Sowemimo-Coker SO, Kirn A, Tribble E, Brandwein HJ, Wenz B. White cell subsets in apheresis and filtered platelet concentrates. Transfusion 1998;38:650–7. [8] Seghatchian J, Krailadsiri P. Validation of different enrichment strategies for analysis of leucocyte sub-population: development and application of a new approach, based on leucofiltration. Transfus Apher Sci 2002;26:61– 72. [9] Rider JR, Want EJ, Winter MA, Turton JR, Pamphilon DH, Nobes P. Differential leucocyte subpopulation analysis of leucodepleted red cell products. Transf Med 2000;10:49–58. [10] Mascaretti L, Baggi L, Riva M, Proserpio P, Dassi M, Varallo F, et al. Lymphocyte subsets in inline packed red blood cell units: comparison between low and high spin procedures. Tranfus Apher Sci 2002;26:167–74. [11] Triulzi DJ, Meyer EM, Donnenberg AD. WBC subsets analysis of WBC-reduced platelet concentrates. Transfusion 2000;40:771–80. [12] Barclay GR, Walker B, Gibson J, McColl K, Turner ML. Quality assurance by a commercial flow cytometry method of leucodepletion of whole blood donations: initial application of universal testing and proposals for a batch-release sampling plan. Transf Med 2000;10:37–48. [13] Steneker I, Beiwenga J. Histologic and immunohistochemical studies on the preparation of white cell-poor red cell concentrates: the filtration process using three different polyester filters. Transfusion 1991;31:40–6.
H. Trindade et al. / Transfusion and Apheresis Science 28 (2003) 319–327 [14] Sivakumaran M, Saunders P, Hanmer D, Drury M, Seghatchian J. Leucocyte fragments in prestorage-leucodepleted blood products: a qualitative analysis. Transfus Sci 2000;22:81–2.
327
[15] Seghatchian J, Krailadsiri P, Beard M, Bukasa A, Bashir S. Studies on the characterization of the cause of leucoreduction failures with particular reference to extra gatal events. Transfus Apher Sci 2002;26:47–60.