This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Feller, N. Articles by Schuurhuis, G. J. Search for Related Content PubMed PubMed Citation Articles by Feller, N. Articles by Schuurhuis, G. J.

Immunologic Purging of Autologous Peripheral Blood Stem Cell Products Based on CD34 and CD133 Expression Can Be Effectively and Safely Applied in Half of the Acute Myeloid Leukemia Patients

Authors' Affiliation: Department of Hematology, VU University Medical Center, Amsterdam, the Netherlands

Requests for reprints: Gerrit J. Schuurhuis, Department of Hematology, BR 248, VU University Medical Center, De Boelelaan 1117, 1081 HV Amsterdam, the Netherlands. Phone: 31-20-444-3838; Fax: 31-20-444-2601; E-mail: gj.schuurhuis{at}vumc.nl.

	Abstract

Top Abstract Methods Results Discussion References

 
Purpose: Several studies have shown survival benefit by autologous stem cell transplantation in acute myeloid leukemia (AML) after purging of grafts. This has, however, not been confirmed in randomized studies due to high toxicity of purging modalities for normal progenitor/stem cells. In this study, we investigated whether positive selection for CD34⁺ and/or CD133⁺ cells, which results in high recovery of normal progenitor/stem cells, is applicable for purging AML grafts.

Experimental Design: Positive selections of normal stem cells using CD34 and/or CD133 can be done if one or both markers are absent or have dim expression and remain so during the course of the disease. Marker expressions in newly diagnosed AML were measured with flow cytometry using a cutoff value for positivity of 1%. Stability of marker expression was studied by pairwise comparison of material at diagnosis and relapse. Leukemia associated phenotype expression was used to measure the efficacy of tumor cell reduction.

Results: In newly diagnosed AML (n = 165), we found no CD34 and/or CD133 expression in 32% of the cases and dim expression in 20% of the cases. No increase in the percentage of CD34⁺ cells (n = 44) and CD133⁺ cells (n = 29) was found in corresponding relapses. Positive selection using grafts contaminated with AML blasts, showing either no or dim expression of CD34 or CD133, resulted in a 3 to 4 log tumor cell reduction (n = 11) with median 50% recovery of normal stem cells.

Conclusions: Purging by positive selection of CD34⁺ and/or CD133⁺ cells can safely, effectively, and reproducibly be applied in about 50% of AML cases.

If purging in AML is used, it is usually done by the application of mafosfamide. A major limitation of the application of mafosfamide is the small therapeutic window (i.e., difference in the chemosensitivity between leukemia and normal cells), potentially leading to a considerable loss of normal progenitor cells in the transplant, resulting in a delayed engraftment. Furthermore, large differences in sensitivity of leukemia blasts among patients lead to highly unpredictable and partly insufficient elimination (5). Also, other chemical or physical purging methods, such as ether lipid and hyperthermia, do not fulfill the 1 to 2 log tumor cell elimination needed for efficient purging. The latter has been calculated by us in a model based on the assessment of MRD contamination in stem cell transplants and the relation with clinical outcome (6).

An attractive approach to purge stem cell transplants is the application of monoclonal antibodies (mAb). Antibodies against myeloid-specific antigens (CD14, CD15, and CD33) have already been used to purge bone marrow from AML patients in morphologic complete remission (7). As with mafosfamide purging, this type of immunologic purging also results in a considerable loss of normal myeloid progenitor cells. Positive selection for markers present in normal progenitor/stem cells, but not present in AML blasts, would potentially lead to high recovery of normal progenitor/stem cells and high efficacy of tumor cell reduction (TCR). Positive selection with CD34 as a primitive marker has been successfully applied in breast cancer, multiple myeloma, and lymphomas (7). Another candidate for positive selection thus far not applied for purging is CD133. CD133 is a marker found both in human progenitor and hematopoietic stem cells. CD133 selected cells successfully engrafted in a fetal sheep transplantation model (8). Also, in the clinical setting, transplantation with CD133-positive cells, like CD34-positive cells, resulted in a fast recovery of mature blood cells (9, 10). Because AML is considered to be a CD34- and CD133-positive malignancy, positive selection for CD34 and CD133 has thus far not been applied. However, because in some of the AML patients no expression of CD34 and/or CD133 has been found (11–13), it can be hypothesized that if one or both of markers are not expressed at diagnosis and remain so at follow-up, MRD-contaminated PBSC transplants might be purged by positive selection for CD34- and/or CD133-positive cells. The cutoff value for expression of CD34 or CD133 is usually based on 20%, which in fact does not accurately identify true CD34- and CD133-negative AML samples. Previously, we found that true CD34-negative AML samples could be identified using a mixture of biological, immunophenotypic, and cytogenetic characteristics (14). A cutoff level of 1% defines the absence of malignant CD34 cells at lower percentage: CD34-positive cells present below 1% are almost exclusively of normal origin. This was also true for CD133 using the same cutoff value of 1% (14). Notably, not only transplants contaminated with AML cells lacking CD34 and/or CD133 expression can be purged but also AML cells having dim expression of one or both of these markers because positive selection procedures exclude cells with low expression.

In this article, we studied the feasibility of positive selection for CD34- and CD133-positive cells as a purging method in AML. Both the efficacy of TCR in autologous PBSC transplants contaminated with AML blasts, which had either no or dim expression of CD34 and/or CD133, and the recovery of normal CD34- and/or CD133-positive cells were established.

	Methods

Top Abstract Methods Results Discussion References

 
Patients. After receiving informed consent, AML blasts from 165 newly diagnosed patients were immunophenotyped and part of the samples were frozen in liquid nitrogen after Ficoll isolation until use. PBSC transplants were obtained either from AML patients and patients with other hematologic malignancies in situation of complete remission or from solid tumors. The clinical characteristics of the 165 AML patients used in this study were as follows: 81 men and 84 women; French American British Classification: M₀ (11 patients), M₁ (16 patients), M₂ (32 patients), M₃ (11 patients), M₄ (25 patients), M₅ (34 patients), M₆ (7 patients), M₇ (1 patient), BAL (1 patient), RAEB-t (20 patients), and unclassified AML (7 patients of whom 3 have secondary AML). Karyotypes according to Grimwade et al. (15) were as follows: 23 patients had favorable aberrations, 10 had poor prognostic features, 96 were classified as intermediate, and 36 patients had unknown karyotype.

CD34 and CD133 expression in acute myeloid leukemia blasts. AML cells obtained at diagnosis were labeled during 15 minutes at room temperature with either CD34-phycoerythrin (Becton Dickinson, San Jose, CA) or CD133-phycoerythrin (Miltenyi, CLB, Amsterdam, the Netherlands) or the isotype immunoglobulin G1 (Becton Dickinson) combined with CD45-peridinyl chlorophyllin (Becton Dickinson) mAbs. Next, cells were lysed with fluorescence-activated cell sorting lysing solution (Becton Dickinson), washed with PBS containing 0.1% bovine serum albumin, and measured on a FACScalibur flow cytometer (Becton Dickinson), which was daily calibrated with calibration beads (Becton Dickinson). CellQuest software (Becton Dickinson) was used for data acquisition and analysis. The expression of CD34- and CD133-positive cells was established in the blast region, characterized by dim expression of CD45 and low side scatter, and divided into three groups: negative, dim, and positive. Positive expression for CD34 was defined as staining of blasts in the third or fourth log quadrant and for CD133 in the third log quadrant. Dim expression was defined as staining of blasts predominantly in the second log quadrant, but showing a higher fluorescence than the isotype control. The staining was considered as negative when <1% of blasts showed higher expression than the isotype control. Normal stem cells showed CD34 expression in the third/fourth quadrant and CD133 expression in the third quadrant (see Results and Fig. 1).

View larger version (34K): [in this window] [in a new window] [Download PPT slide]   Fig. 1. Heterogeneous expression of CD34 and CD133 in AML. Cells were labeled as described in Methods. Blasts were defined by their dim expression of CD45 with low side scatter; expression of CD34 and CD133 was established on these cells. A, AML sample 1 showed no CD34 expression but was positive for CD133; for AML sample 2 the expression pattern was the other way around; and AML sample 3 showed dim expressions of both CD34 and CD133. The expressions of the isotype control immunoglobulin G1 were used to set quadrants for defining negative and (dim) positive expression of CD34 and/or CD133. B, expression of normal CD34- and CD133-positive cells in a control non-AML PBSC transplant. Note that the expression of CD34 and CD133 in these normal stem cells is higher than the dim expression of CD34 or CD133 in AML 3 (A).

Model systems. The efficacy of TCR by positive selection in AML was first established in a model system in which CD45-FITC–prelabeled newly diagnosed CD34- or CD133-negative AML cells were added to non-AML PBSC products (model I). In this model, we used autoMACS technology for positive selection and AML blasts were identified before and after positive selection by their CD45-FITC fluorescence in FL1 channel on a flow cytometer. Theoretically, using this model system, a 5 to 6 log TCR can be achieved in cases when PBSC transplants showed about 1% AML cells and CD34⁺ cells ranging from 0.1% to 1%. In the next model (model II), instead of using CD45-FITC–prelabeled AML blasts for the identification of the AML blasts, we made use of aberrant marker expression present in leukemia cells, the so-called leukemia-associated phenotypes, which are not present or in very low frequencies in normal cells. The application of leukemia-associated phenotypes for MRD detection has been previously described by us in detail (16). Shortly, leukemia-associated phenotypes were established in newly diagnosed AML by an immunophenotypic labeling using a panel of 12 combinations of mAbs each containing four mAbs conjugated with either FITC, phycoerythrin, peridinyl chlorophyllin, or allophycocyanin on a flow cytometer. The mAb combination used were (a) CD3-CD4-CD45-CD34, (b) CD15-CD13-CD45-CD34, (c) CD2-CD56-CD45-CD34, (d) CD5-CD7-CD45-CD34, (e) CD11c-CD11b-CD45-CD34, (f) CD65-CD117-CD45-CD34, (g) CD61-CD33-CD45-CD34, (h) CD71-CD19-CD45-CD34, (i) CD22-CD90-CD45-CD34, (j) CD42b-CD34-CD45-CD14, (k) HLA-DR-CD20-CD45-CD34, and (l) CD34-CD133-CD45-CD38. Based on the results of the immunophenotypic staining of the leukemia blasts, leukemia-associated phenotypes were defined. The defined leukemia-associated phenotypes allow for measurement of the efficacy of TCR because normal cells do not or in very low frequencies express these leukemia-associated phenotypes. In model III, we used CliniMACS instead of autoMACS technology to purge real PBSC products from AML patients; also here the efficacy of TCR was established using leukemia-associated phenotypes. In the last model (model IV), the efficacy of TCR for AML with dim expression of CD34 or CD133 was established. In this case, the efficacy of TCR was determined by combining either CD45-FITC–prelabeled diagnosis AML cells or nonlabeled diagnosis AML cells with leukemia-associated phenotype present into non-AML PBSC products.

P-glycoprotein assay. We previously showed how to establish the origin of CD34-positive cells present in low percentages in bone marrow of newly diagnosed AML (14). In that study, P-glycoprotein activity was the most reliable tool to discriminate normal from malignant CD34-positive cells. CD34-positive cells were documented as normal when P-glycoprotein activity was in the range of P-glycoprotein activities found in normal CD34-positive cells (1.7-3.7), as measured in control bone marrow, whereas the P-glycoprotein activity of the whole CD34-negative AML blasts was outside this range (14). For CD133, this method could also be used: the range of P-glycoprotein activities of normal CD133-positive bone marrow cells varied from 2.2 to 5.5 (14). In addition, even when CD34- or CD133-positive cells showed a P-glycoprotein activity somewhat outside the range of normal CD34- or CD133-positive cells, but at the same time showed a largely different activity compared with the bulk of CD34- or CD133-negative AML cells, these CD34- and/or CD133-positive cells were considered to be normal.

The activity of P-glycoprotein was established using the difference in fluorescence after incubation with the P-glycoprotein probe Syto16 with and without P-glycoprotein modulator PSC833 (17). P-glycoprotein activity was expressed as the ratio of Syto16 fluorescence in the presence of P-glycoprotein modulator PSC833 and Syto16 fluorescence without modulator present, both after subtraction of the fluorescence of the control (cells in medium alone). A ratio of >1.0 thus indicates P-glycoprotein activity.

Statistical analyses. A Spearman's rank correlation test was used to determine the correlation coefficient between CD34 and CD133 expression in newly diagnosed AML and relapsed samples. The Wilcoxon signed-rank test for paired samples was used with a 95% confidence interval to compare differences in percentage and intensity of CD34 and CD133 expression.

	Results

Top Abstract Methods Results Discussion References

 
Analysis of CD34 and CD133 expression in newly diagnosed acute myeloid leukemia. The expression of both CD34 and CD133 in blasts from 165 newly diagnosed AML patients has been investigated using a cutoff level of 1%. Because these CD34-positive cells showed high expression of CD133 (82 ± 10%, n = 18; ref. 14), this cutoff could also be used for CD133. CD34 and CD133 are heterogeneously expressed in AML (Fig. 1). CD34 expression was below 1% in 36 of 165 cases (21.8%) whereas CD133 expression was below 1% in 38 of 165 cases (23%). In 21 of these cases (12.8%) both markers had expression below 1%. Furthermore, in those cases in which >1% CD34 and/or CD133 expression was present, the level of CD34 and CD133 expression varied among AML samples. In 24 of 165 cases (14.5%) AML blasts showed dim expression for CD34, and in 29 of 165 cases (17.6%) dim expression for CD133 (for definition of no, dim, and positive expression, see Methods). In control PBSC products, the normal progenitor/stem cells had a higher level of expression of both CD34 and CD133 expression compared with AML with dim expression of CD34 and/or CD133 (Fig. 1).

Establishment of the nature of CD133- and CD34-positive cells in newly diagnosed acute myeloid leukemia. The number of cases eligible for purging by positive selection was based on the cutoff level of 1% positive CD34 and/or CD133 expression in AML at diagnosis. In 17 AML samples with <1% expression, we studied the nature of these cells using the P-glycoprotein assay (14). Eleven of 17 samples had both CD34 and CD133 expression below 1%; 5 of 17 samples had only CD34 expression below 1%; and 1 of 17 samples had only CD133 expression below 1%. In 15 of the 16 samples with CD34 < 1% (nos. 1-16, Table 1), the CD34-positive cells were of normal origin according to the definitions in Methods. One sample (no. 12) was not conclusive (). In the 11 samples showing CD133 < 1% (no 2-11 and 17), of which 10 samples showed overlapping CD34-positive blasts in low percentages (nos. 2-11), P-glycoprotein activities were also indicative of the normal character of the CD133-positive blasts. In summary, based on P-glycoprotein activity in 16 of 17 patients showing <1% CD34 and/or CD133 expression, the CD34- and/or CD133-positive cells were of normal origin, whereas only one case was not conclusive.

For CD133, similar to CD34, the percentages of CD133-positive cells at diagnosis and relapse were highly correlated ( = 0.574, P = 0.001, n = 29), with no significant increase of the percentage of CD133-positive cells at relapse (P = 0.71). All five samples that showed very low expression at diagnosis still had low or dim expression at relapse. Notably, in 4 of the 25 evaluable samples, CD133 expression decreased from positive (n = 3) or dim (n = 1) to negative. Also similar to CD34, median fluorescence intensities of CD133 at diagnosis and relapse were correlated ( = 0.481, P = 0.02, n = 21). In contrast to CD34, however, median fluorescence intensity was not increased at relapse (P = 0.67), neither in samples that were CD133 negative at diagnosis nor in samples with dim expression or positive for CD133.

Positive selection of acute myeloid leukemia blasts with no malignant CD34 and/or CD133 expression. The efficacy of TCR by positive selection was firstly established in a model system in which freeze-thawed, CD45-FITC–prelabeled CD34-negative blasts were added to non-AML PBSC products (see model I in Methods and Fig. 2). Using this model, we found a 3.5 to 4.6 log TCR in three independent experiments after positive selection for CD34-positive cells of these contaminated PBSC products (Table 2, model I). The purity and recovery of normal CD34-positive cells were 98% to 99% (median, 98%) and 27% to 78% (median, 41%), respectively. Furthermore, we found that the efficacy of TCR was not affected by the degree of tumor contamination as tested with 0.1%, 1%, and 10% AML cells: the TCR was 3.6 log, 3.5 log, and 3.4 log, respectively (not shown).

View larger version (24K): [in this window] [in a new window] [Download PPT slide]   Fig. 2. Purging efficacy by positive selection using CD45-FITC–prelabeled AML blasts. A, AML blasts can thus be distinguished from normal cells by their CD45-FITC fluorescence signal in FL1 channel. Normal cells showed CD34 expression (with no fluorescence in FL1 channel) whereas AML blasts (with fluorescence in FL1 channel) had no CD34 expression at all. B, after positive selection using CD34 microbeads, the normal CD34-positive cells were enriched and 0.01% CD45-FITC–prelabeled AML blasts were found in the positive fraction. C, after positive selection using CD34 microbeads, the CD45-FITC–prelabeled AML blasts were found in the negative fraction. Few normal CD34-positive cells were also found in this fraction.

Next, we studied the efficacy of TCR by applying positive selection using CliniMACS technology in a real AML PBSC product, in which no AML blasts were added but contained real MRD (model III). MRD in PBSC was 0.27%, as measured using the aberrant marker expression of CD45^dimCD15⁺⁺CD4⁺⁺, as established in this AML at diagnosis. The newly diagnosed AML cells had 0.1% CD34 expression. Positive selection with anti-CD34 microbeads using CliniMACS technology resulted in a 2.8 log TCR. Accurate measurement of the log TCR was limited by the number of leukemia-associated phenotype–positive cells measured in the positive fraction. When we artificially increased the MRD from 0.27% MRD to 2.8% MRD by adding cryopreserved/thawed diagnosis AML blasts to the transplant, the efficacy of TCR under these conditions was 4 log (Fig. 3; Table 2, experiment 8).

View larger version (65K): [in this window] [in a new window] [Download PPT slide]   Fig. 3. Purging efficacy using positive selection with anti-CD34-coupled beads on CliniMACS. Purging efficacy of positive selection with anti-CD34-coupled beads on CliniMACS has been established using leukemia-associated phenotypes. A, AML cells at diagnosis were CD34 negative (<1% CD34 expression) and showed an aberrant phenotype characterized by a high expression of both CD15 and CD4. The PBSC transplant of this patient has been purged by positive selection using anti-CD34-coupled beads and the efficacy of TCR was established using the aberrant marker expression. B, PBSC transplant contained 3.6% normal CD34+ cells (top left) and 2.8% AML cells, characterized by CD15 and CD4 expression (bottom left). After positive selection with CD34 microbeads, the recovery of CD34 was 97.4% and the tumor contamination was 0.001% (middle), offering a 4 log TCR. The negative fraction showed only few normal CD34-positive cells (top right) whereas the AML cells can be identified by their CD15 and CD4 expression (bottom right).

However, caution is needed when the level of expression is close to that of normal CD34- and/or CD133-positive cells. An example of such expression of CD133 is shown in Fig. 4A, illustrating a non-AML graft added to CD45-FITC–prelabeled AML blasts with dim expression of CD133. After a selection procedure, CD133-expressing blasts partly elute into the positive fraction (Fig. 4B, left), thereby decreasing the purging efficacy. In theory, these AML cells might elute into the negative fraction when the binding strength between these cells and the column is decreased. Such can be achieved by lowering the antibody-microbead concentration. Indeed, when using a 4-fold lower CD133 microbead concentration, a 0.5 log extra TCR was achieved, whereas with a 16-fold lower CD133-microbead concentration, a 1.5 log extra TCR was accomplished. This occurs at the price of the recovery of normal stem cells: CD133 recovery was 29% compared with the 96% when using undiluted CD133 microbeads (Fig. 4B). In another experiment, we found similar results for grafts contaminated with AML cells having CD34 dim expression: using a 4-fold lower CD34-microbead concentration resulted in 1 log extra TCR with a 2.6-fold lower recovery of normal progenitor/stem cells compared with the undiluted CD34 microbeads (23% versus 60%; data not shown).

View larger version (42K): [in this window] [in a new window] [Download PPT slide]   Fig. 4. Efficacy of TCR using different concentrations of beads in AML blasts with CD133 dim expression. A, PBSC product containing normal CD133-positive cells was combined with AML blasts with dim expression of CD133, prelabeled with CD45-FITC. Note that the levels of CD133 expression of AML blasts and normal CD133 blasts are partly overlapping. Positive selection using different concentrations of anti-CD133 beads was applied to obtain a maximal purging. B, positive fraction after selection with different concentration of CD133 beads resulted in 0.06%, 0.02%, and 0.002% recovery of AML blasts, respectively (TCR was 3.2, 3.7, and 4.7 log, respectively). The recovery of normal CD133-positive cells was 96.3%, 67.1%, and 29.0%, respectively.

Application spectrum of positive selection as a potential purging technique. Next, we documented in how many cases positive selection as purging method can potentially be applied. Table 3 summarizes all possible combinations of marker expressions for CD34 and CD133. In 32.1% of the AML patients at diagnosis, no expression of CD34 and/or CD133 was found. As shown above, in these cases a 3 to 4 log TCR can potentially be achieved by positive selection. In 20% of the cases, we found either a dim expression of both markers or a dim expression of one marker combined with positive expression of the other marker. Also, in these cases positive selection can be done with a 3 to 4 log TCR. As argued earlier, when AML blasts had relatively high median fluorescence intensity of CD34 and/or CD133, which is close to the median fluorescence intensity of normal stem cells, the efficacy of TCR might be negatively affected. In that case, a decrease in the antibody-microbead concentration can be applied: this may result in an increase of the efficacy of TCR up to 1.5 log.

	Discussion

Top Abstract Methods Results Discussion References

 
Purging of autologous PBSC transplants remains a controversially issue. In AML several studies showed an actual advantage of purging for survival (3, 4), but no randomized studies have been done to unequivocally prove the principle. The damage to normal stem cells that usually accompanies TCR, and therefore offers a potential risk for the patient, contributes to the reluctance to introduce purging as a logical component of autologous stem cell transplantation. If this disadvantage would be circumvented, it would make sense to aim for a graft that is as clean as possible. In AML the most common purging technique used is mafosfamide treatment which has variable purging efficacies and is hampered by a relatively high toxicity towards normal progenitor stem/cells, which results in a relatively high morbidity and/or mortality.

Positive selection using CD34-positive cells from grafts, with the aim to transplant only cells relevant for hematologic recovery, has been applied both in solid tumors and hematologic malignancies with high recoveries of CD34-positive cells (7). Positive selection for CD34-positive cells has never been applied in AML because AML was considered to be a CD34-positive stem cell disease; together with purified normal stem cells, also purified malignant stem cells would be reinfused. However, in most cases AML is considered as CD34 negative when using cutoff values of 10% to 20% (11, 13), a level which is, of course, much too high to identify true CD34-negative AML samples. In a previous study, we found that AML samples could be identified as truly CD34 negative when using a cutoff level of 1%: the low percentage of CD34-positive cells present in these samples has been proven to be of normal origin (14). Using the strict cutoff value of 1% CD34 expression, we found that 21.8% of all AML samples were truly CD34 negative (Table 3). In these cases, CD34-positive selection can thus be applied without risk of contamination with CD34-positive leukemia cells. For CD133, similar arguments can be used. It is known that CD133-positive normal blasts have the same potential for hematologic recovery as CD34-positive cells (9, 10). When including CD133, the number of cases where positive selection can potentially be applied increases to 32.1% (Table 3).

From experience with positive selection procedures of bone marrow samples, which, in contrast to PBSCs, contain CD34-positive cells with high and dim expression, we knew that cells with dim CD34 expression are eluted in the negative fraction. It can thus be expected that AML cells with dim expression of CD34 and/or CD133 will also be eluted in the negative fraction. Indeed, this seemed to be the case with about similar TCR efficacies as found for AML lacking CD34 and/or CD133 expression. However, when expression patterns of CD34- or CD133-positive AML cells partly overlap with normal cells, this results in contamination with AML of the positive fraction. For these cases, changes of technical conditions, resulting in decreased binding of cells to the column, might result in preferential elution of cells with dim marker expression into the negative fraction. Such conditions can be achieved by either increasing the flow speed during the selection procedure or by lowering the concentration of antibody microbeads. The second approach, used in this study, indeed resulted in a higher TCR, although at the expense of the normal stem cell/progenitor recovery. Nevertheless, the yield of normal stem cells under the latter conditions is still relatively high compared with that using mafosfamide (5% residual clonogenic ability left; ref. 3). When AML samples with no and dim CD34 and/or CD133 expression are taken together, the number of transplants that are potentially eligible for successful purging increases from 32% to 52%.

One potential pitfall of our approach might be that CD34 and/or CD133 expression increases in the period between diagnosis and stem cell collection. Indication for this comes from the finding that CD34 expression was higher at relapse compared with diagnosis (11, 13). We did not find such an increase in a group of 44 paired samples tested. A large difference between our study and other studies is the cutoff value to discriminate between negative and positive expression. Our cutoff level of 1% enabled identification of true CD34-negative AML. The higher cutoff levels of 10% and 20% expression, usually taken in studies, easily allow small increases of marker expressions to cross the border definition for positivity. This is especially true when the expressions of the markers are not established in a blast gate, characterized by CD45 dim expression and a low side scatter. Furthermore, caution is needed for selection procedures in AML that had dim expression of either CD34 or CD133 or both at diagnosis, because in these cases the median fluorescence intensity of CD34 may increase at relapse. It has to be remarked that no data are available on the intensity of expressions under the conditions of stem cell harvest, whereas a down-regulation of CD117 has been described both for leukemia and normal stem cells (6, 18). In case of up-regulation of these markers in PBSC transplants, positive selection using these markers will not be suitable for purging purposes. Therefore, in cases of dim expression of CD34 and/or CD133 in AML at diagnosis, it would be wise to investigate the nature of the CD34- and/or CD133-positive cells present in the autologous PBSC transplants. This can be done using immunophenotypic aberrancies, usually present in CD34- and/or CD133-positive AML blasts at diagnosis (16): positive selection can safely be done in the absence of the aberrant phenotypes in the stem cell transplant. This was indeed the case as tested in three PBSC transplants from AML patients who showed at diagnosis no or dim expression of CD34- and CD133-positive cells.

Positive selection using CD34 has been applied in breast cancer, multiple myeloma, and lymphomas (7). Thus far, most studies have been focused on the improvement of the purging efficacy and the safety of this application (7). No benefit for survival has been published yet. The reasons for this might be that purging is superfluous in some of the transplants, on the one hand, and the emergence of relapses from the endogenous tumor burden, on the other hand. Therefore, the benefits of purging can be only proven in large, randomized studies or in randomized studies with patient groups that have a particular MRD level that is known to predict poor prognosis (e.g., 0.05% as described by us previously; ref. 6).

In conclusion, in the present study we report that positive selection using CD34 or CD133 antibodies coupled to microbeads results in a consistent 3 to 4 log tumor cell reduction in PBSC grafts of AML patients who at diagnosis showed no or dim expression of CD34 and/or CD133. These marker expressions are routinely measured in AML blasts at diagnosis as support for the classification of the leukemia. The high and reproducible recovery of normal progenitor/stem cells, which is common using selections with the CliniMACS, avoids the clinical problems potentially associated with other purging modalities. At present, this approach is already feasible in about half of the patients. If other targets for positive selection would be identified, positive selection might include all AML patients. Such may include not only positive selections of normal stem cells but also negative depletions of the transplant using markers that are expressed in AML but not in normal stem cells.

	Acknowledgments

 
We thank Miltenyi Biotec for the material support.

	Footnotes

 
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received 1/ 6/05; revised 3/22/05; accepted 4/ 1/05.

	References

Top Abstract Methods Results Discussion References

 

1. Burnett AK, Goldstone AH, Stevens RM, et al. Randomised comparison of addition of autologous bone-marrow transplantation to intensive chemotherapy for acute myeloid leukaemia in first remission: results of MRC AML 10 trial. UK Medical Research Council Adult and Children's Leukaemia Working Parties. Lancet 1998;351:700–8.[CrossRef][Medline]
2. Brenner MK, Rill DR, Moen RC, et al. Gene-marking to trace origin of relapse after autologous bone-marrow transplantation. Lancet 1993;341:85–6.[CrossRef][Medline]
3. Gorin NC, Aegerter P, Auvert B, et al. Autologous bone marrow transplantation for acute myelocytic leukemia in first remission: a European survey of the role of marrow purging. Blood 1990;75:1606–14.[Abstract/Free Full Text]
4. Miller CB, Rowlings PA, Zhang MJ, et al. The effect of graft purging with 4-hydroperoxycyclophosphamide in autologous bone marrow transplantation for acute myelogenous leukemia. Exp Hematol 2001;29:1336–46.[CrossRef][Medline]
5. Miller CB, Zehnbauer BA, Piantadosi S, Rowley SD, Jones RJ. Correlation of occult clonogenic leukemia drug sensitivity with relapse after autologous bone marrow transplantation. Blood 1991;78:1125–31.[Abstract/Free Full Text]
6. Feller N, Jansen-van der Weide MC, van der Pol MA, Westra AH, Ossenkoppele GJ, Schuurhuis GJ. Purging of peripheral blood stem cell transplants in AML: a predictive model based on minimal residual disease burden. Exp Hematol 2005;33:120–30.[Medline]
7. Berkahn L. Immunologic methods of purging in autologous stem cell transplantation. J Hematother Stem Cell Res 2000;9:147–59.[Medline]
8. Yin AH, Miraglia S, Zanjani ED, et al. AC133, a novel marker for human hematopoietic stem and progenitor cells. Blood 1997;90:5002–12.[Abstract/Free Full Text]
9. Lang P, Bader P, Schumm M, et al. Transplantation of a combination of CD133+ and CD34+ selected progenitor cells from alternative donors. Br J Haematol 2004;124:72–9.[CrossRef][Medline]
10. Gordon PR, Leimig T, Babarin-Dorner A, et al. Large-scale isolation of CD133+ progenitor cells from G-CSF mobilized peripheral blood stem cells. Bone Marrow Transplant 2003;31:17–22.[CrossRef][Medline]
11. Baer MR, Stewart CC, Dodge RK, et al. High frequency of immunophenotype changes in acute myeloid leukemia at relapse: implications for residual disease detection (Cancer and Leukemia Group B Study 8361). Blood 2001;97:3574–80.[Abstract/Free Full Text]
12. Wuchter C, Ratei R, Spahn G, et al. Impact of CD133 (AC133) and CD90 expression analysis for acute leukemia immunophenotyping. Haematologica 2001;86:154–61.[Abstract/Free Full Text]
13. Thomas X, Campos L, Archimbaud E, et al. Surface marker expression in acute myeloid leukaemia at first relapse. Br J Haematol 1992;81:40–4.[Medline]
14. van der Pol MA, Feller N, Roseboom M, et al. Assessment of the normal or leukemic nature of CD34+ cells in acute myeloid leukemia with low percentages of CD34 cells. Haematologica 2003;88:983–93.[Abstract/Free Full Text]
15. Grimwade D, Walker H, Oliver F, et al. The importance of diagnostic cytogenetics on outcome in AML: analysis of 1,612 patients entered into the MRC AML 10 trial. The Medical Research Council Adult and Children's Leukaemia Working Parties. Blood 1998;92:2322–33.[Abstract/Free Full Text]
16. Feller N, van der Pol MA, van Stijn A, et al. MRD parameters using immunophenotypic detection methods are highly reliable in predicting survival in acute myeloid leukaemia. Leukemia 2004;18:1380–90.[CrossRef][Medline]
17. van der Pol MA, Broxterman HJ, Westra G, Ossenkoppele GJ, Schuurhuis GJ. Novel multiparameter flow cytometry assay using Syto16 for the simultaneous detection of early apoptosis and apoptosis-corrected P-glycoprotein function in clinical samples. Cytometry 2003;55B:14–21.[CrossRef]
18. To LB, Haylock DN, Dowse T, et al. A comparative study of the phenotype and proliferative capacity of peripheral blood (PB) CD34+ cells mobilized by four different protocols and those of steady-phase PB and bone marrow CD34+ cells. Blood 1994;84:2330–939.

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Feller, N. Articles by Schuurhuis, G. J. Search for Related Content PubMed PubMed Citation Articles by Feller, N. Articles by Schuurhuis, G. J.