International Journal of Hematology

, 90:292

CD34+ cell subpopulations detected by 8-color flow cytometry in bone marrow and in peripheral blood stem cell collections: application for MRD detection in leukemia patients

Authors

    • Department of PathologyKarolinska University Hospital, Solna and Karolinska Institute
  • Astrid Gruber
    • Department of Medicine, Division of HematologyKarolinska University Hospital
  • Joanna Mazur
    • Department of Child and Adolescent HealthNational Research Institute of Mother and Child
  • Anna Mårtensson
    • Department of Clinical Immunology and Transfusion MedicineKarolinska University Hospital
    • Databyrån för informationsbehandling
  • Mona Hansson
    • Department of Clinical Immunology and Transfusion MedicineKarolinska University Hospital
  • Anna Porwit
    • Department of PathologyKarolinska University Hospital, Solna and Karolinska Institute
Original Article

DOI: 10.1007/s12185-009-0389-z

Cite this article as:
Björklund, E., Gruber, A., Mazur, J. et al. Int J Hematol (2009) 90: 292. doi:10.1007/s12185-009-0389-z

Abstract

Fast development in polychromatic flow cytometry (PFC) makes it possible to study CD34+ cells with two scatter and eight fluorescence parameters. Minimal residual disease (MRD) is determined as persistence of leukemic cells at submicroscopic levels in bone marrow (BM) of patients in complete remission. MRD can be present in collections of hematopoietic stem cell from blood (HSC-B). Using PFC, we have defined patterns of antigen expression in CD34+ cell subpopulations in BM and applied them as templates in MRD analysis. Twelve BM samples from hospital control (HC) patients with no signs of hematological malignancy were studied using five 8-color monoclonal antibody combinations detecting subsets of CD34+ cells. These patterns have been used as templates to determine levels of MRD in HSC-B collections from six AML patients. Several subsets of CD34+ precursor cells were found to be present at very low frequencies (<10−4) in BM and/or HSC-B collections. All six HSC-B collections from AML patients showed MRD by 8-color technique and only three by previously applied 3-color method. The 8-color technique showed promising results in efficient detection of different CD34+ subpopulations of HSC-B and in MRD quantification. Monitoring of MRD should become a part of quality control of HSC-B collections.

Keywords

Hematopoietic stem cell transplantationMinimal residual diseaseAcute myeloid leukemiaCD34+ cellsFlow cytometry

1 Introduction

Identification and comparison of CD34+ cells and their subpopulations from different sources have been done extensively during the past 2 decades. Hematopoietic stem cells (HSC) have been phenotypically characterized by flow cytometry (FC) in normal peripheral blood (PB), bone marrow (BM), cord blood (CB), and peripheral blood stem cell collections (HSC-B) [113]. Studies of CD34+ cells have mainly been done using 3- or 4-color FC techniques and most of these studies have focused on BM samples.

Fast development in polychromatic FC (PFC) has made it possible to study the CD34+ cells with two scatter and eight monoclonal antibodies (MAbs) conjugated with eight different fluorochromes. This new technique has not yet been well documented, but it may improve the possibility to define and separate CD34+ cell subpopulations [14, 15]. Immunophenotypic characteristics of the normal CD34+ cell subsets are of great interest for detection of minimal residual disease (MRD).

FC studies of MRD are based on the presence of phenotypic features in blast cells that are absent in normal BM cells [1620]. Normal CD34+ patterns are used as templates for MRD studies, since the CD34 antigen is expressed in leukemic blasts in at least 60% of acute myeloid leukemia (AML) patients [21]. The knowledge that some antigens are absent or rarely expressed on normal CD34+ cells is very useful when tailoring MAbs combinations for MRD follow-up in BM samples from leukemia patients. The new PFC technique could allow more efficient detection and more accurate MRD quantification. It is also of great interest to determine if the patterns of the CD34+ cell subpopulations in HSC-B collections are similar to those in the BM. Studies of MRD in HSC-B collections are rather scarce in literature [2225]. MRD studies could be performed as a part of the quality control process.

The first aim was to establish a standard 8-color panel with PFC and to define the patterns of antigen expression in human CD34+ precursor cell subpopulations from non-leukemic BM. This panel, based on antibody combinations previously applied in our MRD studies [26], served to identify phenotypes that can be used as templates in MRD analysis. Moreover, we wanted to determine if the CD34+ subset patterns in reactive BM differ from those in HSC-B collections from patients with CD34-negative hematological diseases. The obtained information has been applied to establish 8-color FC MRD analysis in HSC-B collections from AML patients. The aberrant cell populations were detected more readily with 8-color analysis in comparison to previously performed MRD studies with 3-color FC.

2 Materials and methods

2.1 Patients

Six adults and six children without hematological diseases formed a hospital control (HC) group for establishing reference patterns for CD34+ subpopulations in the BM. The adult group had a median age of 55 years (range 27–64 years, M:F 2:4) and the children group had a median age of 10 years (range 3 months–16 years, M:F 2:4). Six patients had reactive bone marrow and were investigated, due to fever of unknown origin, reactive lymphadenopathy, suspected temporal arthritis or idiopathic thrombocytopenic purpura (ITP), to exclude lymphoma or leukemia. In four patients, BM samples were taken for staging of lymphomas or other tumors, but showed no BM involvement by both morphology and FC. In two children, BM samples were taken 6 months after the end of treatment for B-precursor ALL. These samples showed no signs of MRD by extensive FC analysis [27].

Eight HSC-B collections were obtained from patients with CD34-negative hematological disease (M:F 6:2, median age 50 years, range 32–57 years). This group included four patients with non-Hodgkin’s lymphoma (NHL), three patients with multiple myeloma (MM) and one patient with plasma cell leukemia.

Six HSC-B collections from AML patients (M:F 2:4, median age 46, range 39–52 years) were also investigated. The primary AML diagnosis was made between 1995 and 2002, based on morphology and cytochemistry following the French–American–British (FAB) classification [28]. Two AML samples were classified as M0, three as M2 and one as M4. All samples were reviewed and reclassified according to the World Health Organization (WHO) classification [29].There were one AML with t (8;21) translocation and five AML not otherwise categorized (2 minimally differentiated AML, 2 AML with maturation and 1 acute myelomonocytic leukemia). The phenotypes at diagnosis are given in Table 1.
Table 1

MRD levels in HSC-B collections from AML patients with 3- and 8-color technique

No

Diagnosis FAB

3-Color FC technique

8-Color technique

Outcome

MAb comba

MRD% in HSC-B collection

MAb combb

MRD% in HSC-B collection

1

M4

CD2-117-34

0.04

No 4

0.10

CR

2

M2

CD34-7-19

0.06

No 5

0.06

CR

3

M2

CD65-34-33

<0.1

No 4

0.02

R, BMT

4

M2

CD33-13-38-34

<0.1

No 2

0.01

CR

5

M0

CD7-2-34

0.02

No 4

0.02

R/D

6

M0

CD33-13-34

<0.1

No 2

0.02

R, BMT

HSC-B Hematopoietic stem cells in blood, BMT bone marrow transplantation, CR complete remission, R relapse, D death, DR HLADR, s subpopulation

a3-color technique combinations were tailored depending on phenotypes at diagnosis. Pat no1: positive for CD34(s), CD15, CD13, CD2(s), TdT, DR, dimCD45, negative for CD117 and CD33; pat no2: positive for CD34, CD7, CD33, CD13, CD19, dimCD45, negative for CD117; pat no 3: positive for CD34, CD65, CD13, strongDR, CD38, dimCD45, negative for CD117 and CD33; pat no 4: positive for CD34, CD13, CD15, CD123, DR, CD38, dimCD4(s), dimCD45, negative for CD33 and 117; pat no 5: positive for CD34, CD117, CD7, CD13, CD15(s), DR, CD38, TdT, dimCD45, CD2(s), negative for CD33; pat no 6: positive for CD34, CD13, DR(s) dimCD4, dimCD45, negative for CD33 and CD117

bThe best combination detecting MRD with the standard 8-color MAb combination panel (Table 2)

2.2 Sample collection and processing

2.2.1 FC studies on CD34+ cells in reactive BM

FC studies were performed on heparinized BM aspirates. The BM samples were diluted 1:1(v/v) ratio with 0.9% NaCl solution and kept at room temperature until processed (within 24 h after collection).

2.2.2 FC studies on CD34+ cells in HSC-B collections

HSC-B cells were mobilized after 3 or 4 cycles of conventional chemotherapy and administration of 10 μg/kg rhG-CSF (Hoffman la Roche, Basel, Switzerland). Collections were performed according to standard procedures using COBE BCT Spectra blood separators (COBE BCT Inc, Lakewood, CO, USA). Briefly, collections were initiated on the day when CD34+ cell levels exceeded 20 × 106/L and continued daily until a target of >3 × 106 CD34+ cells/kg was achieved (in most patients >5 × 106/kg). The cells were frozen in autologous plasma with 10% DMSO (dimethyl sulphoxide, BDH, Merck AG, Dietikon, Switzerland) in a controlled-rate freezer (Planer Kryo 10 series 3, Sunbury on Thames, UK).

2.3 Antibody panel, data acquisition and analysis

Aliquots of HSC-B collections were thawed rapidly, washed and re-suspended in PBS containing 10% fetal calf serum. Cells were diluted to a concentration of 10 × 106/ml.

As much as 100 μl of sample was incubated for 10 min in a dark at room temperature with five different MAb combinations as listed in Table 2. The design of the 8-color panel was based on a previous standard 3-color panel (CD34-CD133-HLADR, CD34-CD90-HLADR, CD34-CD38-HLADR, CD34-CD117-CD14, CD15-CD33-CD34, CD34-CD7-CD19, CD61-CD34-CD13, TdT-CD34-CD45, CD45RA-CD45RO-CD34) used from previous studies for the assessment of immunophenotype of CD34+ cells in HSC-B collections. The 3-color panel was tailored according to present literature and the knowledge of the most frequent aberrances in AML immunophenotyping.
Table 2

Five 8-color MAb combinations used to delineate CD34 + cell subsets

Combination

FITC

PE

PerCP5.5a or PE-Cy5b

PE-Cy7

APC

APC-Cy7

PA

AM

1

CD38

CD90

CD34b

CD33

CD133

HLADR

CD45RA

CD45

2

CD13

CD133

CD33a

CD117

CD34

HLADR

CD4

CD45

3

CD38

CD45RO

CD34b

CD13

CD117

HLADR

CD45RA

CD45

4

CD15

CD45RO

CD33a

CD34

CD117

CD14

CD45RA

CD45

5

TdT

CD7

CD61a

CD13

CD34

HLADR

CD19

CD45

MAbs directly conjugated with fluorescein isothiocyanate (FITC), phycoerythrin (PE), peridin chlorophyll protein 5.5 (PerCP5.5) or PE-cyanine 5 (PE-Cy5), PE-cyanine 7 (PE-Cy7), allophycocyanin (APC), allophycocyanin 7 (APC-Cy7), Pacific Blue (PA) and AmCyan (AM). Antibody source: CD133PE and CD133APC from MACS, Milteyi Biotec, Germany; CD13F and CD38F from Dako, Glostrup, Denmark; CD7PE, CD34 PECy5, CD45ROPE and CD117APC from Immunotech, Marseille, France; CD117PECy-7, CD19PA and CD45RAPA from Biolegand, San Diego, USA and all other antibodies from Becton Dickinson (BD), San Jose, CA, USA

The incubation was followed by a lysing step (5 min) with 2 ml FACS lysing solution (BD, diluted 1:10 in distilled water). After lysis, the cells were centrifuged (5 min, 500g), washed in 2 ml FACS flow (BD) and re-suspended in 0.5 ml of FACS flow. A similar stain and lyse/wash method was used for BM samples [27]. The intracellular markers were analyzed after fixation and permeabilization using Intra-Stain (DAKO, Glostrup, Denmark).

2.3.1 Data acquisition

Acquisition was performed using either FACScan (for 3-color FC) with Cell-Quest or FACS-Canto II with the Diva software program [all from Becton Dickinson (BD), San Jose, CA, US]. Instrument setup, calibration, and quality instrument control were performed using commercial standard reagents Spherocaliflow QC kit (Spherotech, Lybertyville, IL, US) or BD cytometer setup and tracking beads (BD), following standard rules [30]. Briefly, the compensation matrix is automatically calculated within the software in Canto II. Lot number-specific calculations are routinely performed for all tandem-conjugated MAbs, and CST beads are daily used to check the stability of the fluorescence detectors in the flow cytometry.

2.3.2 Data analysis

Data analysis was performed using Infinicyt software (Cytognos, Salamanca, Spain). A gating strategy applied in order to determine immunophenotypes of CD34+ cell subpopulations is illustrated in Fig. 1, and the populations of very low frequency are marked in bold in Suppl. Fig 1.
https://static-content.springer.com/image/art%3A10.1007%2Fs12185-009-0389-z/MediaObjects/12185_2009_389_Fig1_HTML.gif
Fig. 1

a Analysis of CD34+ cell subsets in bone marrow from an adult hospital control patient Representative dot-plots of the analysis of CD34+ subsets with the MAb combination 4 (Table 2) are shown. Upper row The analysis started by excluding debris and dead cells in FSC/SSC dot-plot. CD34+ BM cells (fulfilling the criteria of CD34+, CD45RO+ dim, CD45RA+ dim and FSC/SSC low/intermediate scatter) are displayed as violet dots, 386 violet events and 250000 totally acquired cells gave 0.15% CD34+ cells in the total BM. Lower row Two subsets of CD34+ cells are shown: CD33+/CD15neg (372 blue events) and CD33neg/CD15neg (14 orange events). There were no CD34+ populations co-expressing CD15+/CD33+ and CD15+/CD33neg cells. All the CD34+/CD33+/CD15neg cells also expressed CD45RA+ dim/CD45RO+ dim (372 green events). In this population, there were CD117 +/CD14neg cells (330 red events) and CD117neg/CD14neg cells (35 cyan events). No populations of CD34+ cells that were CD117neg/CD14+ or CD117+/CD14+ were found (less than 10 events). All found subpopulations are finally shown in the FSC/SSC dot-plot. The main population: CD34+/CD33+/CD15neg/CD45RA+ dim/CD45RO+ dim/CD117+/CD14neg is shown in red events. The frequencies of the various CD34+ subsets for combination 4 in BM samples are shown in Suppl. Table 1. b Analysis of CD34+ cell subsets in HSC-B collection from a patient with CD34-negative hematological disease. Representative dot-plots of the analysis of CD34+ subsets with the MAb combination 4 (Table 2) are shown. The gating strategy was applied as defined in a, upper row. CD34+ cells were 1.6% (4115 events/250000 cells). Upper row Displays two main subsets of CD34+ cells: CD33+/CD15neg (1034 blue events) and CD15+/CD33+ (3072 violet events). No clear populations of CD33neg/CD15neg and CD15+/CD33neg cells were found (less than 10 events). The two CD34+ CD33+ populations with different CD15 expression were both CD45RA/CD45RO-positive. One population dimly expressed and the other (CD34+ CD33+ CD15+) with higher expression of CD45RA/CD45RO. These populations are also illustrated in SSC/CD34 and FSC/SSC dot-plot. Lower row The CD34+/CD33+/CD15+/CD45RA+/CD45RO+ population also expressed CD14 and CD117 (2836 cyan events). These cells were smaller and had lower side scatter. The subpopulation CD34+/CD33+/CD15+/CD45RA+ dim/CD45RO+ dim were CD117+/CD14neg (866 red events) and showed higher intensity of CD34 expression. These cells showed also higher FSC

2.4 MRD studies

When 3-color MRD studies were performed, individual follow-up protocols based on immunophenotyping results from the leukemia diagnosis were tailored and used to determine the MRD levels in BM remission samples and in HSC-B collections (Table 1). In MRD studies with the 8-color technique, the standard five-combination panel (Table 2) was used in all samples. The acquisition procedure continued until all cells in the tube had passed through the flow cytometer (at least 250000). During the analysis, the percentages of events with characteristics corresponding to leukemia-related immunophenotypes were evaluated. Clusters of at least 10–15 cells with leukemia-related phenotype and scatter characteristics were considered as detectable MRD. In each tube, a total of at least 250000 cells passed through the flow cytometer, corresponding to a sensitivity level of 10−4 (detection of 0.01% leukemic cells) [21, 27, 31].

The procedures were in accordance with the ethical standards of the institutional committee on human experimentation and the 1975 Helsinki Declarations as revised in 2000.

2.5 Statistical methods

Descriptive statistical values including mean, median, standard deviation (SD) and range were used [32].

3 Results

3.1 CD34+ subset patterns in reactive BM samples

To establish the reference patterns for CD34+ subpopulations in the BM, 12 reactive BM samples from HC patients were analyzed using the antibody panel shown in Table 2. The median percentage of CD34+ cells was 0.45% (range 0.23–0.46%) in the adult HC group. In children, the median percentage of CD34+ cells was 2.0% (range 0.89–6%). The typical patterns found in the five antibody combinations are shown in Suppl. Fig 1 and Table 1. The CD34+ cells displayed a heterogeneous light scatter distribution. Almost all (>95%) CD34+ cells expressed HLADR, CD38, CD45RAdim, and CD45ROdim. On average, 80% of CD34+ cells were also positive for CD133, CD90, CD13, and CD33. Most (>70%) of the CD34+ cells were CD117 and dimly CD4-positive. No clear CD34+ cell subsets also displaying CD14, CD7, and CD61 were found. The CD34+/CD19+ subpopulation was very low in adults (mean 0.02%, SD 0.026). In children, a considerable subpopulation of CD19+ CD34+ was present (mean 0.46, SD 0.38). Also, the CD34+ TdT+ subpopulation was higher in children (mean 0.56%, SD 0.56) compared to adult BM (mean 0.05%, SD 0.10). In children, there were also more CD34+ cells expressing HLADR+/CD133+/CD38+/CD45RA+/CD90+, but negative for CD33. Details on differences between reactive BM from adult and children are given in Suppl. Table 1.

3.2 CD34+ subset patterns in HSC-B collections

The median percentage of CD34+ cells was 0.63% (range 0.57–0.96) in eight HSC-B collections from patients with CD34-negative hematological diseases. CD34+ cells displayed a heterogeneous light scatter distribution together with a dim CD45 expression. Most patterns of antigen expression in CD34+ subsets from HSC-B collections were similar to those in BM samples (Suppl. Table 1). However, a CD34+ population co-expressing CD15 and CD33 was noticed in all eight HSC-B collections and could not be found in the BM samples. This population was also seen in six HSC-B collections from AML patients and comprised at least 10% of the CD34-positive cells. The CD34+/CD33+/CD15+ cells also expressed CD117 and CD14 (Fig. 1).

3.3 MRD studies

MRD was previously studied in HSC-B collection of six AML patients with 3-color standard FC panel. The phenotypes at diagnosis and applied 3-color MAb combinations are given in Table 1. In 3-color analysis, three out of six studied HSC-B collections were MRD-positive (Table 1). Samples from the same six collections were investigated using the panel with five 8-color MAb combinations (Table 2). Patterns of CD34+ subsets in the BM samples as well as from HSC-B collections from patients with CD34-negative hematological diseases were used as reference (Table 3; Suppl. Fig 1, Table 1). With 8-color technique, we found aberrant populations with phenotypes suggesting leukemic cells in all six collections from AML patients (Table 1, Fig. 2). All six AML patients received their collections. Three patients are in continuous complete remission. The other three suffered leukemia recurrence. One of these patients died of leukemia. Two other patients received allogeneic BM transplant and are in continuous remission (Table 1).
Table 3

Frequencies of various CD34 subpopulations in HSC-B performed with a standard 8-color panel

Comb.1

Total CD34+

CD34+DR+CD133+

CD34+DR+CD133neg

CD34+DRnegCD133neg

CD34+DRneg’CD133+

CD34+DR+CD133+CD38+CD45RAd+

CD34+DR+CD133+CD38+CD45RAneg

Non AML

N = 8

2.2 (0.62, 3.2)

[0.1–7.43]

2.2 (0.55, 3.1)

[0.08–8.1]

0.1 (0.03, 0.18)

[0–0.64]

<0.01

no cluster

<0.01

no cluster

1.89 (0.53, 2.7)

[0.08–6.45]

<0.01

no cluster

AML

N = 6

0.35 (0.25, 0.30)

[0.07–0.83]

0.34 (0.20, 0.28)

[0.06–0.85]

0.01 (0.01, 0.01)

[0–0.03]

<0.01

no cluster

<0.01

no cluster

0.32 (0.23, 0.29)

[0.06–0.78]

<0.01

no cluster

Comb.2

Total CD34+

CD34+DR+CD133+

CD34+DR+CD133neg

CD34+DRnegCD133neg

CD34+DRnegCD133+

CD34+DR+CD133+CD13+CD33+

CD34+DR+CD133+CD13+CD33neg

Non AML

N = 8

2.3 (0.57, 3.4)

[0.11–8.34]

2.2 (0.55, 3.1)

[0.08–8.1]

0.1 (0.03, 0.18)

[0–0.64]

<0.01

no cluster

<0.01

no cluster

0.33 (0.33, 0.09)

[0.2–0.45]

0.07 (0.07, 0.04)

[0.02–0.12]

AML

N = 6

0.37 (0.21, 0.32)

[0.11–0.94]

0.34 (0.20, 0.28)

[0.06–0.85]

0.01 (0.01, 0.01)

[0–0.03]

<0.01

no cluster

<0.01

no cluster

1.8 (0.33, 2.8)

[0.09–7.3]

0.45 (0.17, 0.67)

[0.01–2]

Comb.3

Total CD34+

CD34+DR+CD38+

CD34+DRnegCD38+

CD34+DRnegCD38neg

CD34+DR+CD38neg

CD34+DR+CD38+CD45ROd+CD45RAd+

CD34+DR+CD38+CD45ROd+CD45RAneg

Non AML

N = 8

2.37 (0.63, 3.5)

[0.08–9.4]

2.3 (0.58, 3.5)

[0.08–9.35]

<0.01

no cluster

<0.01

no cluster

<0.01

no cluster

2.0 (0.54, 3.0)

[0.07–8.06]

<0.01

no cluster

AML

N = 6

0.26 (0.24, 0.17)

[0.05–0.5]

0.25 (0.23, 0.17)

[0.05–0.49]

<0.01

no cluster

<0.01

no cluster

<0.01

no cluster

0.19 (0.16, 0.13)

[0.05–0.41]

<0.01

no cluster

Comb.4

Total CD34+

CD34+CD15negCD33+

CD34+15negCD33neg

CD34+CD15+CD33+

CD34+CD15+CD33neg

CD34+CD15neg

CD133+CD45ROd+CD45RAd+

CD34+CD15negCD38negCD45RO+dCD45Raneg

Non AML

N = 8

1.8 (0.68, 2.1)

[0.39–5.6]

1.0 (0.29, 1.5)

[0.01–4.1]

0.23 (0.11, 0.29)

[0.02–0.99]

0.31 (0.13, 0.36)

[0.02–0.99]

0.07 (0.04, 0.06)

[0.01–0.16]

0.99 (0.28, 1.5)

[0.01–13.9]

<0.01

no cluster

AML

N = 6

0.36 (0.29, 0.25)

[0.09–0.78]

0.18 (0.17, 0.11)

[0.06–0.35]

0.015 (0.005, 0.027)

[0–0.07]

0.13 (0.13, 0.11)

[0.01–0.3]

0.02 (0.02, 0.02)

[0–0.06]

0.17 (0.16, 0.11)

[0.05–0.33]

<0.01

no cluster

Comb.5

Total CD34+

CD34+CD13+DR+

CD34+CD13negDR+

CD34+CD13negDRneg

CD34+CD13+DRneg

CD34+CD13+DR+CD7+CD61neg

CD34+CD13+DR+CD7negCD61neg

Non AML

N = 8

2.1 (1.0, 1.9)

[0.5–5.4]

1.7 (0.92, 1.6)

[0.4–4.6]

0.37 (0.12, 0.42)

[0.09–1.2]

<0.01

no cluster

<0.01

no cluster

<0.01

no cluster

1.7 (0.9, 1.5)

[0.39–4.5]

AML

N = 6

0.74 (0.53, 0.55)

[0.15–1.7]

0.47 (0.40, 0.34)

[0.06–0.98]

0.23 (0.12, 0.28)

[0.07–0.8]

<0.01

no cluster

<0.01

no cluster

<0.01

no cluster

0.46 (0.40, 0.33)

[0.06–0.97]

Comb.1

Total CD34+

CD34+DR+CD133+CD38negCD45RAneg

CD34+DR+CD133+CD38negCD45RAd+

CD34+DR+CD133+CD38+CD45RAdimCD90d+CD33+

CD34+DR+CD133+CD38+CD45RAdimCD90d+CD33neg

CD34+DR+CD133+CD38+CD45RAdimCD90negCD33neg

CD34+DR+CD133+CD38+CD45RAdimCD90negCD33+

Non AML

N = 8

2.2 (0.62, 3.2)

[0.1–7.43]

<0.01

no cluster

<0.01

no cluster

1.7 (0.4, 2.5)

[0.08–6]

<0.05

no cluster

<0.01

no cluster

<0.01

no cluster

AML

N = 6

0.35 (0.25, 0.30)

[0.07–0.83]

<0.01

no cluster

<0.01

no cluster

0.29 (0.22, 0.26)

[0.06–0.72]

<0.05

no cluster

<0.01

no cluster

<0.01

no cluster

Comb.2

Total CD34+

CD34+DR+CD133+CD13negCD33neg

CD34+DR+CD133+CD13negCD33+

CD34+DR+CD133+CD13+CD33+CD4d+CD117+

CD34+DR+CD133+CD13+CD33+CD4negCD117+

CD34+DR+CD133+CD13+CD33+CD4negCD117neg

CD34+DR+CD133+CD13+CD33+CD4d+CD117neg

Non AML

N = 8

2.3 (0.57, 3.4)

[0.11–8.34]

<0.01

no cluster

<0.01

no cluster

1.2 (0.27, 1.9)

[0.06–5,2]

0.38 (0.06, 0.60)

[0.02–1.4]

0.04 (0.08, 0.005)

[0–0.23]

0.11 (0.27, 0.18)

[0.01–0.51]

AML

N = 6

0.37 (0.21, 0.32)

[0.11–0.94]

<0.01

no cluster

<0.01

no cluster

0.20 (0.16, 0.16)

[0.05–0.47]

0.05 (0.03, 0.06)

[0.01–0.18]

<0.05

no cluster

<0.05

no cluster

Comb.3

Total CD34+

CD34+DR+CD38+CD45RonegCD45RAneg

CD34+DR+CD38+CD45ROneg+CD45RAd+

CD34+DR+CD38+CD45ROd+45RAd+CD13+CD117+

CD34+DR+CD38+CD45ROd+45RAd+CD13+CD117neg

CD34+DR+CD38+CD45ROd+45RAd+CD13negCD117neg

CD34+DR+CD38+CD45ROd+45RAd+CD13negCD117+

Non AML

N = 8

2.37 (0.63, 3.5)

[0.08–9.4]

<0.01

no cluster

<0.01

no cluster

1.4 (0.42, 2.0)

[0.05–4.9]

0.54 (0.1, 1.0)

[0.02–3.0]

<0.01

no cluster

<0.01

no cluster

AML

N = 6

0.26 (0.24, 0.17)

[0.05–0.5]

<0.01

no cluster

<0.01

no cluster

0.13 (0.10, 0.1)

[0.03–0.26]

0.04 (0.01, 0.06)

[0–0.015]

<0.01

no cluster

<0.01

no cluster

Comb.4

Total CD34+

CD34+CD15+CD33+CD45RonegCD45RAneg

CD34+CD15+CD33negCD45ROneg+CD45RAd+

CD34+CD15negCD33+CD45ROd+45RAd+CD14negCD117+

CD34+CD15negCD33+CD45ROd+45RAd+CD14negCD117neg

CD34+CD15negCD33+CD45ROd+45RAd+CD14d+CD117neg

CD34+CD15negCD33+CD45ROd+45RAd+CD14d+CD117+

Non AML

N = 8

1.8 (0.68, 2.1)

[0.39–5.6]

<0.01

no cluster

<0.01

no cluster

1.0 (0.27, 1.4)

[0.01–3.8]

0.04 (0.21, 0.07)

[0.08–0.37]

<0.01

no cluster

0.32 (0.28, 0.14)

[0.11–0.58]

AML

N = 6

0.36 (0.29, 0.25)

[0.09–0.78]

<0.01

no cluster

<0.01

no cluster

0.15 (0.16, 0.08)

[0.05–0.25]

<0.01

no cluster

<0.01

no cluster

0.17 (0.16, 0.09)

[0.07–0.31]

Comb.5

Total CD34+

CD34+CD13+DR+CD7negCD61+

CD34+CD13+DR+CD7+CD61+

CD34+CD13+DR+CD7negCD61negCD19negTdTneg/d+

CD34+CD13+DR+CD7negCD61negCD19+TdTneg/d+

CD34+CD13+DR+CD7negCD61negCD19+TdT+

CD34+CD13+DR+CD7negCD61negCD19negTdT+

Non AML

N = 8

2.1 (1.0, 1.9)

[0.5–5.4]

<0.01

no cluster

<0.01

no cluster

1.7 (0.9, 1.5)

[0.38–4.44]

<0.01

no cluster

<0.01

no cluster

<0.01

no cluster

AML

N = 6

0.74 (0.53, 0.55)

[0.15–1.7]

<0.01

no cluster

<0.01

no cluster

0.45 (0.39, 0.33)

[0.06–0.96]

<0.01

no cluster

<0.01

no cluster

<0.01

no cluster

All values in percentage (%) of total cells and presented as: mean (median, SD) [range]

https://static-content.springer.com/image/art%3A10.1007%2Fs12185-009-0389-z/MediaObjects/12185_2009_389_Fig2_HTML.gif
Fig. 2

MRD analysis in HSC-B collection from an AML patient. Leukemic cells at diagnosis expressed CD34, CD13, HLADR (subpopulation), dim CD4 and dim CD45, and lacked expression of CD33 and CD117 (patient 6, Table 1). Representative dot-plots of the analysis of CD34+ subsets with combination 2 (Table 2) are shown. The gating strategy was applied as in Fig. 1a, upper row. Totally, there were 0.7% CD34+ cells (1752 events/250000 acquired cells) in the HSC-B collection. Upper row Two CD34+ cells subsets are shown: CD13+/CD33neg (203 blue events) and CD13+/CD33+ (755 violet events). The population CD13+/CD33neg was also CD117neg/HLADR+ (31 red events) and displayed CD4dim+. Totally, 209844 acquired cells gave an MRD value of 0.015% aberrant leukemic cells in the collection. The populations are also illustrated in the SSC/CD34+ dot-plot. Lower row Analysis of HSC-B collection from a CD34-negative hematological disease patient used as control. Representative dot-plot histograms of the analysis of CD34+ subsets with combination 2 (Table 2) are shown. The gating strategy is defined in Fig. 1a, Upper row. Two subsets of CD34+ cells are shown: CD13+/CD33neg (32 blue events) and CD13+/CD33+ (920 violet events). In the CD13+/CD33neg population there were less than 10 events with the leukemia-associated phenotype (CD117neg/HLADR+/CD4dim+). The populations are also illustrated in the SSC/CD34+ dot plot

4 Discussion

To establish reference values for 8-color FC technique, we defined the patterns of antigen expression and frequency of various subsets of CD34+ cells in reactive BM. Our results confirm previous publications, which reported that scatter distribution of CD34+ cells and their immunophenotypes are heterogeneous.

Except for the populations of CD34+/CD33+/CD15+ cells found in HSC-collections, the patterns of antigen expression in reactive CD34+ cells from BM and HSC-collections were similar. It has previously been reported that the distribution of different CD34+ subsets in HSC-B collection might significantly vary depending on the applied mobilization regimen [6, 10]. Although there were some differences in mobilization regimes applied to patients with AML and those with non-hematological diseases in our study, there were no differences when immunophenotypic profiles of CD34+ cells from HSC-B were compared. Reported differences in expression of various antigens in CD34+ subsets may also depend on the MAb clones and fluorochrome conjugates applied in various studies [33]. In all groups of patients in this study, we used the same clones, fluorochromes and the same five 8-color combinations.

Previous studies have reported various features of the CD34+ cell phenotypes in HSC-B collections [6, 33]. Our results confirm that the multipotent stem cells defined by the absence of CD38 and HLADR on their surface represent less than <0.01% of all PB-mobilized CD34+ cells [33, 34]. We have also confirmed that phenotypes such as CD34pos/HLADRpos/CD38neg and CD34pos/CD38pos/HLADRneg had very low frequency (<0.01% of total BM cells) in the reactive BM. Due to normally low frequencies, these phenotypes could be applied as MRD markers if found in AML patients.

D’Arena et al. [4] had reported that the pluripotent CD34+ cells expressed CD45RO. Our study showed that >95% of the CD34+ cells expressed CD45RAdim CD45ROdim, and >80% also expressed CD133. The 5-transmembrane receptor-CD133 has previously been described by different groups [35, 36]. Goussetis et al. showed that a subpopulation of CD34high cells expressed CD133. They proposed that CD34high CD133+ cells may represent early progenitors with high proliferation rate, while the subsets of CD34+ CD133dim and CD34+ CD133 cells may represent late-committed progenitors with limited proliferative potential. In the present study, we could not distinguish the CD34/CD133 subsets in different subpopulation by intensity.

In both BM and HSC-B collections, we could confirm that the great majority (>80%) of CD34+/CD133+/HLADR+ cells co-expressed CD33 and CD13 antigen. We have determined the frequency of CD34+/CD33−/CD13+ (mean 0.07–0.39% of total BM cells) and showed that the subpopulation of CD34+/CD33+ CD13 cells and CD34+/HLADR+/CD133+ cells, which lack both CD33 and CD13 expression, was even more uncommon (<0.01% of the total BM cells). The lack of expression of either CD13 or CD33 represents common aberrant phenotypical features in AML. Therefore, these phenotypes could be applied as MRD markers in AML patients. It has been previously suggested that either these two markers appeared simultaneously during the very early stages of myeloid differentiation [16], or that the expression of CD13 preceded that of CD33 during myeloid differentiation and that CD34+/CD13+/CD33neg cells corresponded to the earlier stages of myeloid cell differentiation [9]. In our study, the CD34+ cells expressing only CD13 had smaller size and lower SSC than those expressing only CD33, which supports the findings by Gaipa et al. [9].

A recently published study defined immunophenotypes in three categories of CD34+ cells: immature CD34+ cells, CD34+ neutrophil precursors, and CD34+ B cell precursors. The immature CD34+ precursor cells and the CD34+ neutrophil precursor cells in BM expressed both CD13 and CD33. CD13 and CD33 had different MFI values in these populations, which may be related to the different stages of maturation [13]. We have shown that the subset CD34+/CD38+/HLADR+/CD33+ CD13+ also expressed markers associated with early hematopoietic cells (CD90, CD133 and CD117). Using the residual mature lymphocytes and neutrophils as reference, we found that these cells expressed CD4 dimly. We confirmed that patterns of antigen expression in CD34+ cells from reactive BM and HSC-B collections were similar to some part. However, in all HSC-B collections, small populations of CD34+ cells positive for CD15/CD33 were noticed (median 0.13% of total BM cells Fig. 1), which were absent in normal BM. In comparison to the whole CD34+ population, these cells showed a higher expression of CD45RA and CD45RO, a lower expression of CD34 and were positive for CD14 and CD117. These cells may represent monocytes or myeloid dendritic cell precursors [37].

In agreement with previous publications from other groups, the other lineage-specific markers were not found in high numbers of CD34+ cells, both in the adult BM and HSC-B collections in our study [16, 38]. Co-expression of CD34 and lymphoid-associated markers TdT, CD19, and CD7 are rather frequently detected as aberrant markers in AML blasts at diagnosis, and can be used for MRD detection. In normal adult BM and HSC-B collections, the CD34+ cells expressing CD19, TdT or CD7 represented <0.1% of total BM cells. However, in children we noticed higher levels of CD34+/CD19+ and CD34/TdT+ cells. As reported previously, there were also higher total numbers of CD19+ ells in the children’s BM (mean 9.9%) in comparison to adults (mean 2.1%) [17].

None or only very few CD34+/CD7+ cells were reported by some authors in normal blood and BM [1, 18], while others found small populations of CD34+/CD7+ cells in the BM [13, 3941]. In our study, CD34+/HLADR+/CD13+/CD7+ cells constituted <0.01% of total BM cells in all studies samples except in the childhood HC group where the mean value was higher.

Some, but not all authors, have also found small subpopulations of CD34+ cells expressing the megakaryocytic lineage-associated antigens [3, 42]. We could not confirm the presence of any CD34+ subpopulation co-expressing CD61.

HSC-B transplantation was introduced due to presumed lower contamination by leukemic blasts as compared to the autologous BMT. However, literature data on MRD levels in HSC-B collections are scarce. MRD detection may be applied as a tool in the quality control of HSC-B grafts. Using PCR-based methodology, Seriu et al. demonstrated the presence of MRD in eight of 13 studied HSC-B collections from children with ALL and suggested that MRD analysis may allow selection of grafts with better quality [43]. In the study of Feller et al. [12], MRD at levels higher than 0.1% was found in HSC-B collections from 69% of 36 evaluated AML patients. Other authors found a strong correlation of pre-transplant MRD detection with subsequent relapse in an autologous transplant setting [44, 45]. Using PFC, CD34+ cells in HSC-B collection from patients with CD34neg hematological diseases showed similar patterns of antigen expression as normal BM samples. Using these patterns as reference, we were able to detect populations with aberrant features in all six studied HSC-B collections from AML patients. These collections had previously been studied by a 3-color technique and MRD could be detected in only three of six samples (Table 1). However, in most samples the frequencies of cells with aberrant phenotypes were very low. Considering the small number of studied patients and follow-up showing that three patients, who received the MRD-positive HSC-B, remained in continuous CR, the clinical significance of these findings is still unclear.

5 Conclusion

The 8-color technique and PFC allowed a better definition of the CD34+ subsets. It was also more efficient in the detection of aberrant populations in HSC-B collections from AML patients in comparison to previously performed MRD studies with 3-color FC analysis. The established patterns of antigen expression by normal CD34+ cells may contribute to the easier and more reliable identification of aberrant phenotypes for MRD investigation. However, for validation of this methodology, a much higher number of comparative analyses with the current gold standard are necessary. Monitoring of MRD could be included as a part of the quality control of HSC-B collection.

Acknowledgments

The excellent technical assistance of Agnieszka Dul, Britt-Marie Johansson, Ana Lodoli, and Kia Heimersson is gratefully acknowledged. We thank Lewis Edgel for linguistic consultation.

Supplementary material

12185_2009_389_MOESM1_ESM.ppt (44 kb)
Supplementary material 1 (PPT 43 kb)
12185_2009_389_MOESM2_ESM.doc (148 kb)
Supplementary material 2 (DOC 148 kb)

Copyright information

© The Japanese Society of Hematology 2009