Flow Cytometric Analysis of Normal B Cell Differentiation

Leukemia (1999) 13, 419–427  1999 Stockton Press All rights reserved 0887-6924/99 $12.00 http://www.stockton-press.co.uk/leu Flow cytometric analysis of normal B cell differentiation: a frame of reference for the detection of minimal residual disease in precursor-B-ALL PLu´cio1, A Parreira1, MWM van den Beemd2, EG van Lochem2, ER van Wering3, E Baars3, A Porwit-MacDonald4, E Bjorklund4, G Gaipa5, A Biondi5, A Orfao6, G Janossy7, JJM van Dongen2 and JF San Miguel6

BIOMED-1 Concerted Action: 1Department of Hematology, Portuguese Institute of Oncology, Lisbon, Portugal; 2Department of Immunology, University Hospital Dijkzigt/Erasmus University Rotterdam, The Netherlands; 3Dutch Childhood Leukemia Study Group, The Hague, The Netherlands; 4Department of Pathology, Karolinska Hospital, Stockholm, Sweden; 5Clinica Pediatrica, University of Milan, Ospedale San Genaro, Monza, Italy; 6Department of Hematology, University of Salamanca, Salamanca, Spain; and 7Department of Immunology, Royal Free Hospital, London, UK

During the last two decades, major progress has been made in Introduction the technology of flow cytometry and in the availability of a large series of monoclonal antibodies against surface membrane and intracellular antigens. Flow cytometric immunophen- Differentiation of B cells from early committed progenitors to otyping has become a diagnostic tool for the analysis of normal mature B-lymphocytes is a multistep maturation process that and malignant leukocytes and it has proven to be a reliable can be monitored by the coordinated acquisition and loss of approach for the investigation of minimal residual disease leukocyte differentiation antigens. Several hypothetical mod- (MRD) in leukemia patients during and after treatment. In order els of B cell differentiation have been proposed in which the to standardize the flow cytometric detection of MRD in acute consecutive sequence of antigen expression is described, thus leukemia, a BIOMED-1 Concerted Action was initiated with the enabling the definition of different stages of B cell participation of six laboratories in five different European coun- 1–3 tries. This European co-operative study included the immuno- maturation. phenotypic characterization and enumeration of different pre- Initially, most concepts of human early B cell development cursor and mature B cell subpopulations in normal bone were derived from studies on precursor-B-acute lymphoblastic marrow (BM). The phenotypic profiles in normal B cell differen- leukemia (precursor-B-ALL).4–6 More recently, studies have tiation may form a frame of reference for the identification of focused on the investigation of normal bone marrow (BM)7–9 aberrant phenotypes of precursor-B cell acute lymphoblastic leukemias (precursor-B-ALL) and may therefore be helpful thanks to technical improvements and refinements of multi- in MRD detection. Thirty-eight normal BM samples were anal- parameter flow cytometry, which allows the analysis of large yzed with five different pre-selected monoclonal antibody numbers of normal B cells, even when present at very low combinations: CD10/CD20/CD19, CD34/CD38/CD19, CD34/ frequencies.10 In addition, extensive panels of monoclonal ؊ CD22/CD19, CD19/CD34/CD45 and TdT/CD10/CD19. Two CD19 antibodies against B cell-associated leukocyte differentiation immature subpopulations which coexpressed B cell-associated antigens have been generated, whole blood/BM staining antigens were identified: CD34؉/CD22؉/CD19؊ and ؉ ؉ ؊ ؎ methods have been developed, new fluorochromes with high TdT /CD10 /CD19 , which represented 0.11 0.09% and 11,12 of the total BM nucleated cells, respectively. sensitivity emission signals have been discovered and 0.05% ؎ 0.04 These immunophenotypes may correspond to the earliest methods for flow cytometric intracellular stainings have been stages of B cell differentiation. In addition to these minor improved substantially.13 These combined technical advances ؉ subpopulations, three major CD19 B cell subpop- have contributed to a better insight in the discrete maturation ulations were identified, representing three consecutive stages of normal B cell development. Several controversies /maturation stages; CD19dim/CD34؉/TdT؉/CD10bright/CD22dim CD45dim/CD38bright/CD20؊ (subpopulation 1), CD19؉/CD34؊/ concerning the consecutive expression of antigens throughout TdT؊/CD10؉/CD22dim/CD45؉/CD38bright/CD20dim (subpopulation B cell differentiation have been clarified, and minor B cell and CD19؉/CD34؊/TdT؊/CD10؊/CD22bright/CD45bright/CD38dim/ subpopulations have been identified in normal BM, such as (2 CD20bright (subpopulation 3). The relative sizes of subpopula- those coexpressing CD34 and CD20 or CD34 and CD22.14–16 tions 1 and 2 were found to be age related: at the age of 15 Nevertheless, some controversies remain: classical models3,5,6 years, the phenotypic precursor-B cell profile in BM changed have been challenged by the possible existence of parallel from the childhood ‘immature’ profile (large subpopulations 1 17 and 2/small subpopulation 3) to the adult ‘mature’ profile (small pathways of B cell development; furthermore, differences subpopulation 1 and 2/large subpopulation 3). When the immu- between fetal and postnatal BM B cell precursors have been nophenotypically defined precursor-B cell subpopulations reported5,17,18 indicating also that postnatal age-related differ- from normal BM samples are projected in fluorescence dot- ences which mainly involve the relative distribution of the dif- plots, templates for the normal B cell differentiation pathways ferent BM B cell subsets, exist.9,19,20 However, in these studies can be defined and so-called ‘empty spaces’ where no cell procedures have been used in which pre-enrichment steps for populations are located become evident. This allows discrimination between normal and malignant precursor-B cells and mononuclear cells are included. This limits the potential value can therefore be used for MRD detection. of the enumeration of the different BM cell subsets due to the Keywords: normal B cell differentiation; minimal residual disease; existence of important levels of uncontrolled loss of cells of precursor-B-ALL; flow cytometry; lymphopoiesis interest.21 Accordingly, to the best of our knowledge, no sys- tematic study of B cell differentiation at different ages has been performed in a large series of healthy individuals using a whole bone marrow sample preparation technique. In the present study, we performed a standardized flow cytometric analysis at six different sites of a total of 38 normal human BM samples from healthy individuals whose age Correspondence: A Orfao, Servicio General de Citometria, Laborato- rio de Hematologia, Hospital Clinico Universitario, Paseo de San ranged from 2 to 75 years. Our major goal was to investigate Vicente, 58–82, 37007 Salamanca, Spain; Fax: 34 923 29 46 24 the BM B cell compartment in order to generate a frame of Received 29 December 1997; accepted 21 October 1998 reference for the identification of leukemia-associated pheno- Flow cytometric analysis of B cell differentiation PLućio et al 420 types in precursor B-ALL cases. The definition of the normal Immunophenotypic analysis patterns of B cell differentiation was based on: (1) the objec- tive identification of clearly defined subsets of B cells corre- For standardization purposes, all groups used the same set of sponding to discrete stages of B cell maturation; (2) the rela- monoclonal antibodies conjugated with fluorescein isothiocy- tive distribution of each of these subsets from the total BM anate (FITC), phycoerythrin (PE) and either peridin chlorophyll cellularity; and (3) the definition of flow cytometry dot-plot protein (PerCP) or the PE/cyanine 5 (PE/Cy5) fluorochrome templates corresponding to the normal B cell differentiation tandem. The antibodies’ clones were carefully selected on the pathways. basis of their reactivity patterns and the absence of back- ground staining. Their source and specificity were as follows: CD10-FITC (W8E7), CD19-FITC (leu12), CD20-PE (leu16), CD22-PE (leu14), CD34-FITC (HPCA2), CD34-PE (HPCA2), Materials and methods CD38-PE (leu17) and CD45-PerCP (Hle1) were purchased from BD, CD10-PE (J5) and CD19-PE/Cy5 (J4.119) were from Specimen collection Coulter/Immunotech (Miami, FL, USA), CD19-PE/Cy5 (SJ25- C1) was from CALTAG Laboratories (San Francisco, CA, USA) BM samples were collected from 38 healthy donors by six and TdT-FITC (HT6) from DAKO (Glostrup, Denmark). participating European immunophenotyping laboratories Subsequently, the triple fluorochrome-conjugated mono- involved in the BIOMED-1 Concerted Action: Investigation of clonal antibody combinations were extensively tested in the Minimal Residual Disease in Acute Leukemia: International six participating laboratories according to predefined guide- Standardization and Clinical Evaluation (Department of Hem- lines. Based on the combined results, the following five triple atology, Portuguese Institute of Oncology, Lisbon, Portugal; immunofluorescence stainings were established: CD10-FITC/ Department of Immunology, University Hospital CD20-PE/CD19-PE-Cy5, CD34-FITC/CD38-PE/CD19-PE-Cy5, Dijkzigt/Erasmus University Rotterdam, Rotterdam, The CD34-FITC/CD22-PE/CD19-PE-Cy5, CD19-FITC/CD34-PE/ Netherlands; Dutch Childhood Leukemia Study Group, The CD45-PerCP, and TdT-FITC/CD10-PE/CD19-PE-Cy5. Irrel- Hague, The Netherlands; Department of Pathology, Karolin- evant monoclonal antibody reagents of the same isotypes as ska Hospital, Stockholm, Sweden; Clinical Pediatrica, Univer- in the triple combinations, conjugated with the same fluor- sity of Milan, Ospedale San Genaro, Monza, Italy; Depart- ochromes, were used as negative controls. ment of Hematology, University of Salamanca, Salamanca, For standardized calibration of the flow cytometer, normal Spain). peripheral blood lymphocytes were stained with CD3-FITC Informed consent was obtained from all donors and parents (Leu-4, BD), CD4-PE (Leu-3, BD), and CD8-PE/Cy5 (3B5; in the case of children less than 15 years in accordance with CALTAG Laboratories) and were measured in parallel with the local or national guidelines of the Medical Ethics Commit- each BM sample. Data acquisition was performed in all cen- tees. The median age of the 38 BM donors was 15 years, rang- ters with FACS can flow cytometers (BD), using either the Lysis ing from 2 to 75. Age ranges for the subjects included in this II (BD) or CellQuest (BD) software programs. In all cases, a study, in each individual center, were as follows: Portuguese two-step acquisition procedure was used. Accordingly, 15 000 Institute of Oncology, 5 to 50 years; Erasmus University of events from total BM were acquired first; in the second step, Rotterdam, 8 to 60; Dutch Childhood Leukemia Study Group, a B cell gate was established on the basis of CD19 expression 2 to 17 years; Karolinska Hospital, 2 to 25; Ospedale San and low right angle light scatter (SSC), as previously Genaro from Monza, 3 to 11; and University of Salamanca, described23 and a minimum of 100 000 events were analyzed 24 to 75 years. All children (n = 18) included in this study for each monoclonal antibody combination. This gate was were BM donors for transplantation. BM samples from adults used to acquire data for the specific characterization of the (Ͼ15 years old; n = 20) were obtained from either BM donors different B cell subpopulations. Additional live gates were for transplantation or healthy subjects undergoing ortho- defined based on CD34 or TdT expression for better discrimi- pedic surgery. nation between the most immature subsets and for identification of potential CD19 negative B cell precursors. The percentage of each subset was calculated from the total number of nucleated BM cells as previously described.24 Sample preparation The analyses were performed on either Hewlett-Packard or MacIntosh computers using the Paint-A-Gate software (BD). BM samples were collected in heparin or EDTA anticoagulant, immediately diluted 1/1 (vol/vol) in phosphate-buffered saline (PBS) and maintained at room temperature until processed. In Standardization of the procedures and quality control all cases, sample preparation and flow cytometric data acquisition were performed within the first 24 h after collection. For All procedures used for the immunophenotypic analysis of the the immunophenotypic studies a stain and then lyse technique normal BM B cell compartment were standardized among the was used. Briefly, 2 × 106 nucleated cells were incubated for six participating centers prior to the study. This included both 10 min with saturating amounts of fluorochrome-conjugated experimental work and regular meetings of scientists and tech- mouse anti-human monoclonal antibodies at room tempera- nicians from all participating centers. For that purpose the ture in the dark, followed by incubation for 10 min with 2 ml same reagents and techniques were used throughout the study of FACS lysing solution (Becton Dickinson (BD), San Jose, CA, at each individual site. As mentioned above decisions on USA) diluted 1/10 (vol/vol) in distilled water. After centrifug- reagent selection were performed at different levels including ation (5 min, 500 g), cells were resuspended in 1 ml PBS with selection of monoclonal antibody clones, fluorochrome con- 0.2% BSA or subjected to intracellular staining for terminal jugates and triple-staining combinations. The combinations deoxynucleotidyl transferase (TdT) according to a previously which were agreed upon were those providing the most described procedure.13,22 powerful discrimination between the subsets under study. In Flow cytometric analysis of B cell differentiation PLućio et al 421 a further step, the procedures used for instrument set-up and A more mature CD19+ B cell subpopulation was pheno- calibration were also comparatively evaluated and a common typically defined by the absence of CD10 and a decreased and standardized method was selected. The scatter character- CD38 expression, as well as by a brighter expression of CD22 istics and the autofluorescence levels of normal peripheral and CD45 (subpopulation 3, Table 1). This subpopulation rep- blood (PB) lymphocytes were used as the reference, based on resented 2.58 ± 1.43% (range 2.32% to 3.10%) of all BM which instrument settings were established. A specific meeting nucleated cells (Figure 2, blue dots). was devoted to educate the personnel of each participating According to these results, Figure 3 summarizes the laboratory in this regard. As described above a common two- sequence of antigen expression by the different B cell subpop- step acquisition procedure was used in all centers and the ulations that were identified in the BM samples. It is note- same applied to data analysis with Paint-A-Gate software. worthy that a high degree of concordance was observed Quality control of the procedures was performed both prior between the relative proportions of each of the three CD19+ to and during the study. For that purpose analysis of a normal B cell subsets identified by the different MAb combinations. PB sample stained for the CD3-FITC (leu4)/CD4-PE An exception to this observation was the relative distribution (leu3)/CD8-PE-Cy5 (3B5) triple combination was performed of subpopulations 2 and 3, when the CD19/CD34/CD45 triple on a daily basis. Additionally, data files were exchanged and combination was used (Table 1). analyzed at different sites and a cross-comparison of the Looking for the relative size of subpopulations 1 to 3 in results obtained, was performed. different individuals, a high degree of variability was observed in the proportion of subpopulation 2, which appeared to be age-related (see below). Variability in the relative size of the Statistical analysis other subpopulations was less pronounced (Table 1).

All results are expressed as mean values with one standard deviation, as well as range. Statistical signiﬁcance of the differences observed was determined with the paired Student’s t- test (P Ͻ 0.05). Distribution of early CD19-hemopoietic progenitors expressing B cell-associated antigens and CD19+ B cell subpopulations in normal BM according to age Results

Identification of early CD19− hemopoietic cell The relative proportions of all precursor B cell subpopulations progenitors expressing B cell-associated antigens in were analyzed according to the donor’s age, which ranged normal BM and phenotypic characterization of from 2 to 75 years. Regarding the most immature CD19− cells, sequential stages of B cell development a higher proportion of cells from population A was observed in childhood BM, as compared to adult BM (mean value of In order to identify the potential earliest stages of B cell differ- 0.15 ± 0.11% and 0.07 ± 0.02% of the childhood and adult entiation in normal BM, we used two triple combinations BM, respectively; P = 0.03). In contrast, when the CD34/CD22/CD19 and TdT/CD10/CD19 with specific gating TdT/CD10/CD19 combination was used, no significant differ- on CD34+ and TdT+ cells. Two immature subpopulations were ence in the proportion of subpopulation B was found between identified by the coexpression of CD34+/CD22+ and children and adults (0.05 ± 0.07% and 0.04 ± 0.03%, respect- TdT+/CD10+ in the absence of CD19, hereafter called subpop- ively; P = 0.65). ulations A and B, respectively. These subpopulations rep- With regard to the relative size of the CD19+ B cell subpop- resented 0.11 ± 0.09% (Figure 1, blue dots) and 0.04 ± 0.05% ulations, the data showed significant changes in subpopula- (Figure 1, red dots) of the total BM, respectively. Nevertheless, tions 1 and 2, but not in subpopulation 3 (Table 1). With most cells coexpressing CD34/CD22 and TdT/CD10 also respect to the proportion of subpopulation 1 and 2, two expressed CD19 (Figure 1, green dots). groups could be defined: children (р15 years; n = 18) and The mean percentage of CD19+ cells from the total number adults (Ͼ15 years; n = 20). A clear decrease in the relative of nucleated cells present in normal BM was 6.46 ± 4.76%. frequency of B cells belonging to these two subpopulations Within the CD19+ B cells, the most immature cells was observed in older individuals, while the relative fre- (subpopulation 1 in Table 1) were immunophenotypically quency of the more mature B cells (population 3) is stable characterized by the expression of CD10bright, CD34 and TdT, throughout age. and were CD19dim, CD22dim, CD38bright, CD45−/dim and Considering the relative size of the CD19+ subpopulations CD20− as examplified in Figure 2 (green dots). This subpopul- according to the total BM nucleated cells, it was observed that ation covers 0.44 ± 0.65% (range 0.01% to 4.06%) of the total the size of subpopulation 1 was larger in children than in BM nucleated cells. adults (ranges for the different monoclonal antibody combi- An intermediate B cell maturation stage was characterized nations from 0.66 ± 0.84% to 0.77 ± 0.70% in children and by coexpression of CD19, CD10, CD22, and CD45 from 0.12 ± 0.12% to 0.18 ± 0.14% in adults, P р 0.03) (Table (subpopulation 2, Table 1). A clear distinction from subpopul- 1). Subpopulation 2 was found to be the largest B cell subpop- ation 1 was observed, based on the loss of TdT and CD34 ulation in children’s BM and to be significantly larger than in and decreased CD10 expression (Figure 2, red dots); the CD45 adults (range for the different monoclonal antibody combi- and CD20 expression is more intense in this subpopulation, nations of 5.17 ± 6.12% to 7.15 ± 7.43% in children and as compared to more immature CD34+ B cell precursors. 0.75 ± 0.51% to 0.86 ± 0.60% in adults, P р 0.01) (Table 1). Interestingly, we did not find changes in the expression of When assessed as a fraction of the total BM cellularity, the CD22 and CD38, which remained dim and bright, respect- number of cells belonging to subpopulation 3 was similar in ively. This subset of B cells represented 3.75 ± 5.75% (range children and adults, although a trend to lower values could 0.05% to 29.59%) of all BM nucleated cells. be observed in BM samples from older donors (Table 1). Flow cytometric analysis of B cell differentiation PLućio et al 422

Figure 1 Flow cytometric analysis of CD19− cells as assessed in CD34 (plots with blue dots) and TdT (plots with red dots) gated cells (see Results for deﬁnition of cell subpopulations). The possible B-lineage commitment of the CD34+ cells is documented by the expression of CD22 (blue cells, left dot plots). The green dots represent precursor-B cell subpopulation 1 as depicted in Figure 2.

Table 1 Relative sizes of the CD19+ precursor-B cell subpopulations in normal human BM of children (р15 years) and adults (Ͼ15 years)

MAB Subpopulation 1a Subpopulation 2a Subpopulation 3a combination (0.44 ± 0.65%) (3.75 ± 5.75%) (2.58 ± 1.43%) CD19dim, CD34+, TdT+, CD10bright, CD19+, CD34−, TdT−, CD10+, CD19+, CD34−, TdT−, CD10−, CD22dim, CD45dim, CD38bright and CD22dim, CD45+, CD38bright CD22bright, CD45bright, CD38dim and CD20− and CD20dim CD20bright

Children Adults P value Children Adults P value Children Adults P value

CD19/CD34/CD45 0.71 ± 0.96 0.17 ± 0.15 0.036 5.17 ± 6.12 0.75 ± 0.51 0.009 4.06 ± 2.65 2.20 ± 0.97 0.013 (0.00–3.86) (0.00–0.50) (0.20–24.04) (0.12–2.18) (0.74–11.40) (1.16–4.85) CD10/CD20/CD19 0.77 ± 0.70 0.12 ± 0.12 0.002 6.93 ± 7.44 0.80 ± 0.59 0.005 2.61 ± 1.26 2.22 ± 0.99 0.325 (0.14–2.15) (0.00–0.63) (0.43–29.59) (0.05–6.18) (1.11–5.88) (0.85–3.97) CD34/CD22/CD19 0.66 ± 0.84 0.14 ± 0.14 0.022 6.94 ± 7.04 0.86 ± 0.60 0.003 2.50 ± 1.29 2.15 ± 0.90 0.355 (0.01–3.51) (0.01–0.51) (0.48–27.74) (0.08–6.29) (1.20–5.44) (0.96–3.81) CD34/CD38/CD19 0.68 ± 0.83 0.16 ± 0.14 0.028 7.15 ± 7.43 0.83 ± 0.52 0.005 2.58 ± 1.39 2.52 ± 1.00 0.879 (0.08–3.32) (0.01–0.45) (0.52–28.98) (0.15–5.07) (0.87–6.09) (1.11–4.50) TdT/CD10/CD19 0.76 ± 0.78 0.18 ± 0.14 0.022 7.09 ± 7.64 0.80 ± 0.62 0.012 2.45 ± 1.00 2.54 ± 1.05 0.862 (0.10–2.92) (0.02–0.42) (1.55–26.95) (0.13–1.68) (1.27–4.48) (1.36–4.10)

aRelative sizes (%) of each precursor-B cell subpopulation in the total nucleated BM cells. Results are expressed as mean ± s.d. and range (in brackets).

Reproducibility of results tutions and we found a high degree of reproducibility among the centers as regards the dot plot templates obtained and the Since the results presented in this paper were pooled from phenotypic characteristics of the different subsets analyzed; those obtained for different individual BM samples in six dif- this also applied to the relative distribution of each subset ferent institutions, we wanted to confirm that the variations whenever the comparison between centers was performed in found with age were real and not dependent on a differential an age-matched way. distribution of individuals/samples or on results among the centers. Therefore, we performed a comparison of the flow cytometric data files obtained in the six participating insti- Flow cytometric analysis of B cell differentiation PLućio et al 423

Figure 2 Flow cytometric analysis of subpopulations 1, 2 and 3 within the CD19+ gate. Subpopulation 1 (green dots), is CD10bright and CD20− (a), CD45−/dim (c), CD38bright(d), TdT+ (f), CD34+ (e), CD22+ (e). Subpopulation 2 (red dots) is defined by the acquisition of CD20 (a), the loss of CD34 and TdT (c to f) and the decrease of CD10 expression which occur simultaneously with the increase in intensity of CD19. Dot plot ‘b’ illustrates the distinction between subpopulations 1 and 2 when using the CD10/CD19 combination. Subpopulation 3 (blue dots) is characterized by the expression of CD20bright, CD10− CD22bright, CD45bright, and CD38−/dim (dot plots a to f). Flow cytometric analysis of B cell differentiation PLućio et al 424 aspects of the flow cytometric procedures in order to avoid methodological variations. The antibody combinations were carefully selected for the hybridoma clone and fluorochrome conjugations. Also the technical details concerning immunofluorescence staining protocols and flow cytometric analysis were standardized and controlled between all groups. According to our data, the sequence of antigen expression during B cell maturation somehow differs from the classical models of B cell development proposed by Loken et al.1 We observed that the CD22 antigen is already expressed on the membrane of the most immature B cells, preceding CD10 and CD19 expression. This finding contrasts with earlier reports, where the expression of surface CD22 was considered to be a relatively late phenomenon in B cell maturation,1 but is in agreement with more recent studies with sensitive multiparameter flow cytometric analyses.14,31,32 Our results show the existence of immature CD19− BM precursors that coex- press the CD22 B cell associated antigen together with CD34; these cells correspond to 0.15% of the total BM cellularity in Figure 3 Tentative diagram of the sequence of antigen expression children and 0.07% in adults. Regarding the potential B cell in B cell subpopulations 1 to 3 based on the information obtained commitment of these progenitor cells, it should be noted that from 38 normal BM samples. Further investigations are still needed although CD22 is a pan-B antigen which is usually negative to confirm the potential B cell origin of the most immature CD19− + + + + in T, NK and granulomonocytic cells it has been reported as subpopulations A (CD34 /CD22 ) and B (TdT /CD10 ). positive in certain subsets of normal myeloid cells such as mast cells33 and dendritic cells.34 However, through the use Discussion of additional multiple stainings we were not able to confirm the existence of cells coexpressing CD22 and CD34 in the The sequential expression of specific antigens during matu- absence of CD19 that displayed clear reactivity for other den- 33,35 ration of normal precursor-B cells in BM has been extensively dritic cell or mast cell associated markers. studied during the past decades, leading to several models of Using the TdT/CD10/CD19 triple combination, we ident- B cell lineage maturation. Although there is a consensus for ified a very small subpopulation showing reactivity for TdT most antigens, frequently discordant findings are and CD10 in the absence of CD19 (subpopulation B). This is reported.14,25–29 These discrepancies might be due to differ- in line with the earlier observation that expression of CD10 ences in methodological approaches, the origin of hematopo- may precede that of CD19 in early B cell differen- ietic cell samples (fetal liver, fetal BM, or post-natal BM), and tiation17,30,36,37 and that B cells may originate together with T, + + − in the age of BM donors. Furthermore, it is not yet clear NK and dendritic cells from a CD34 /CD10 /lineage precur- whether the sequential expression of functional antigens dur- sor cell.38 Therefore, this subpopulation might represent the ing the early phases of B cell differentiation follows a rigid earliest B cell precursors. Nevertheless,it must be noted that + − sequence or, alternatively, may vary both in normal and in the commitment to the B cell lineage of these CD10 /CD19 malignant B cell development.30 cells is not universally accepted. In a recent study it has been The study presented here was performed in six different lab- shown that CD34+/CD10+/CD19− BM precursor cells give rise oratories. Therefore, careful attention was paid to the stan- to macrophage colonies.39 In the present study we could dardization and quality control of the various technical identify this minor subpopulation in all samples but, in con-

Figure 4 Distribution of B cell subpopulations in normal BM according to age (in years), expressed as a percentage of CD19+ cells. The age-related increase of the relative size of subpopulation 3 is evident in all the analyzed combinations as examplified here for the CD34- FITC/CD38-PE/CD19-PE-Cy5. The transition from ‘childhood’ to ‘adult’ phenotypic patterns appears to occur around the age of 15 years. Flow cytometric analysis of B cell differentiation PLućio et al 425 trast to the other maturation stages, we did not observe a sig- shown). Additional studies are necessary to elucidate the real nificant variation of its relative size with the age of the donors. ontogeny of these two minor populations of BM cells and We also observed that these cells express CD10 at lower den- therefore the potential existence of parallel pathways of B sity than the more mature B cell precursors. Further multiple cell development.17 stainings demonstrated that these cells are CD22− (data not Using five different triple monoclonal antibody stainings, it

Figure 5 Examples of abnormal distributions of leukemic B cells (depicted in red) localizing into the ‘empty spaces’ left by normal precursor- B cells (represented in grey) for each of the five monoclonal antibody combinations used. Flow cytometric analysis of B cell differentiation PLućio et al 426 was possible to discriminate three major distinct CD19+ B cell siglio Nazionale delle Ricerche (Italy); and grants CICYT SAF maturation stages (CD10bright/CD22dim/CD20−/CD38bright/ 94–308 and AECC-95 (Spain). CD45dim, CD10dim/CD22dim/CD20dim/bright/CD38bright/CD45+ and CD10−/CD22bright/CD20bright/CD38dim/−/CD45bright, subpopulations 1, 2 and 3, respectively). Each of the antibody combinations was, by itself, sufficient to identify the three different References stages (Figure 4). Interestingly, slight variations were observed in the relative distribution of populations 2 and 3 with the 1 Loken MR, Shah VO, Dattilio KL, Civin CI. Flow cytometric analysis of human bone marrow. II. Normal B lymphoid development. CD19/CD34/CD45 combination as compared with the other Blood 1987; 70: 1316–1324. multiple stainings. This probably reflects a lower discriminat- 2 LeBien TW, VillaBlanca JG. Ontogeny of normal human B cell ive power of CD45 to clearly separate between populations and T cell precursors and its relation to leukemogenesis. Hematol 2 and 3. In this sense, it should be noted that selection of the Oncol Clin North Am 1990; 4: 835–847. fluorochrome conjugates was essential to obtain the highest 3 Ryan D, Kossover S, Mitchell S, Frantz C, Hennessy L, Cohen H. discriminative power between the different BM cell subsets Subpopulations of common acute lymphoblastic leukemia antigen-positive lymphoid cells in normal bone marrow identified by for each combination. hematopoietic differentiation antigens. Blood 1986; 68: 417–425. As previously shown, childhood BM is enriched for imma- 4 Gobbi M, Caligaris-Cappio F, Janossy G. Normal equivalent cells ture B cells.5,9 Here we observe that the more immature of B cells malignancies: analysis with monoclonal antibodies. Br CD19+ B cell subpopulations 1 and 2 make up to 70% of all J Haematol 1983; 54: 393–403. CD19+ B cells in BM of children under 15 years, a pattern 5 Nadler LM, Korsmeyer SJ, Anderson KC, Boyd AW, Slaughenhoupt that contrasts with that of adults, where subpopulation 3 pre- B, Park E, Jensen J, Coral F, Mayer RJ, Sallen SE, Ritz J, Schlossman + SF. B cell origin of non-T cell acute lymphoblastic leukemia. A dominates, representing over 70% of the total CD19 B cells. model for discrete stages of neoplastic and normal pre-B cell dif- The transition between these two phenotypic patterns of pre- ferentiation. J Clin Invest 1984; 74: 332–340. cursor-B cell subpopulations occurs in a rather abrupt way 6 Anderson KC, Bates MP, Slaughenhoupt B, Pinkus GS, Schlossman around the age of 15 years, mainly due to the relative SF, Nadler LM. Expression of human B cell-associated antigens on decrease of subpopulations 1 and 2 beyond this age. This leukemias and lymphomas: a model of human B cell differen- decrease in subpopulations 1 and 2 beyond the age of 15 is tiation. Blood 1984; 63: 1424–1433. + 7 Loken MR, Shah VO, Hollander Z, Civin CI. Flow cytometric responsible for the overall decrease of the pool of CD19 BM analysis of normal B lymphoid development. Pathol Immuno- 9 cells in adults also found by Caldwell et al. In this sense it is pathol Res 1988; 7: 357–370. not reasonable to think that these differences could be related 8 Civin CI, Loken MR. Cell surface antigens on human marrow cells: to the potential contamination of the BM specimens with PB dissection of hematopoietic development using monoclonal anti- since children usually show higher numbers of PB mature B bodies and multiparameter flow cytometry. Int J Cell Clon 1987; cells as compared to adults. In spite of this, in specific cases 5: 267–288. 9 Caldwell CW, Poje E, Helikson MA. B cell precursors in normal significant levels of contamination of BM samples with PB pediatric bone marrow. Am J Clin Pathol 1991; 95: 816–823. may exist. 10 Terstappen LWMM, Loken MR. 5-Dimensional flow cytometry as The flow cytometric multiparameter analysis presented here a new approach for blood and bone marrow differentials. Cyto- can be used for studying normal B cell maturation but it is metry 1988; 9: 548–556. conceivable that this approach may facilitate the diagnostic 11 Rothe G, Schmitz G for the Working Group on Flow Cytometry and follow-up studies of patients with precursor-B-ALL. For and Image Analysis. Consensus protocol for the flow cytometric immunophenotyping of hematopoietic malignancies. Leukemia instance, it might be important to investigate whether the age- 1996; 10: 877–895. related switch in immunophenotype reported here also occurs 12 Haugland RP. Handbook of Fluorescent Probes and Research in children and adults with precursor-B-ALL. Furthermore, the Chemicals. Molecular Probes: Junction City, OR, 1985. precise identification of normal precursor-B cell subpopula- 13 Groeneveld K, Marvelde JG, Van den Beemd MWM, Hooijkaas tions in BM also permits the definition of ‘templates’ in the H, Van Dongen JJM. Flow cytometric detection of intracellular multidimensional dot-plot representations of each mono- antigens for immunophenotyping of normal and malignant leukocytes. Leukemia 1996; 10: 1383–1389. clonal antibody combination (Figure 5). These dot plot ‘tem- 14 Smith RG, Kitchens SR. Phenotypic heterogeneity of TdT + cells plates’ could not only be considered as a reference frame for in the blood and bone marrow: implications for surveillance of pinpointing of aberrant phenotypes falling into the so-called residual leukemia. Blood 1989; 74: 312–319. ‘empty-spaces’, but also for the detection of an abnormal dis- 15 Saeland S, Duvert V, Pandrau D, Caux C, Durand I, Wrighton N, tribution of the populations found in the normal differentiation Wideman J, Lee F, Banchereau J. Interleukin 7 induces the prolifer- pathways. This approach will most probably be of utmost ation of normal human B cell precursors. Blood 1991; 78: 2229–2238. importance for improving the standardization of strategies and 16 Hurwitz CA, Gore SD, Stone KD, Civin CI. Flow cytometric detec- simplifying the methodology aimed at the detection of MRD tion of rare normal human marrow cells with immunophenotypes in precursor-B-ALL.27,28,40–43 characteristic of acute lymphoblastic leukemia cells. Leukemia 1992; 6: 233–239. 17 Uckun FM. Regulation of human B ontogeny. Blood 1990; 76: Acknowledgements 1908–1923. 18 LeBien TW, Wormann B, Villablanca JG, Law C-L, Steinberg LM, Shah VO, Loken MR. Multiparameter flow cytometric analysis of This work was supported by grants from the European human fetal bone marrow B cells. Leukemia 1990; 4: 354–358. BIOMED-1 Concerted Action entitled Investigation of Minimal 19 Caldwell CW, Patterson WP. Relationship between T200 antigen Residual Disease in Acute Leukemia: International Standardis- expression and stages of B cell differentiation in resurgent hyper- ation and Clinical Evaluation (BMH-CMT 94-1675); Liga Por- plasia of bone marrow. Blood 1987; 70: 1165–1172. tuguesa contra o Cancro (Portugal); The Dutch Cancer Society 20 Dworzak MN, Fritsch G, Fleicher C, Printz D, Froschl G, Buch- inger P, Mann G, Gadner H. Multiparameter phenotype mapping Koningin Wilhelmina Fonds, grant EUR 94–852 (The of normal and post-chemotherapy B-lymphopoiesis in pediatric Netherlands); Cancerfonden (Sweden); Fondazione Tettam- bone marrow. Leukemia 1997; 11: 1266–1273. anti, Associazione Italiana per la Ricerca sul Cancro and Con- 21 Fritsch G, Printz D, Stimpfl M, Dworzak MN, Witt V, Potschger Flow cytometric analysis of B cell differentiation PLućio et al

+ 427 U, Buchinger P. Quantification of CD34 cells: comparison of 33 Escribano L, Orfao A, Diaz-Agustin B, Villarrubia J, Cervero C, methods. Transfusion 1997; 37: 775–784. Lopez A, Garcia-Marcos MA, Bellas C, Fernandez-Canadas S, 22 Horvatinovich J, Sparks S, Borowitz M. Detection of terminal Cuevas M, Sanchez A, Velasco JL, Navarro JL, San Miguel JF. deoxynucleotidyl transferase by flow cytometry: a three-color Indolent systemic mast cell disease in adults: immunophenotypic method. Cytometry 1994; 18: 228–230. characterization of bone marrow mast cells and its diagnostic 23 Macedo A, Orfao A, Ciudad J, Gonzalez M, Vidriales B, Lopez- implications. Blood 1998; 91: 2731–2736. Berges MC, Martinez A, Landolfi C, Can˜izo C, San Miguel JF. 34 Almeida J, Bueno C, Alguero MC, Tabernero MD, Rasillo A, Perez- Phenotypic analysis of CD34 subpopulations in normal human Simon JA, Menendez P, San Miguel JF, Orfao A. Caracterizacion bone marrow and its application for the detection of minimal inmunofenotipica de celulas dendriticas de sangre periferica nor- residual disease. Leukemia 1995; 9: 1896–1901. mal por citometria de flujo. Haematologica 1998; 83 (Suppl. 2): 24 Orfao A, Escribano L, Villarrubia J, Velasco JL, Cervero C, Ciudad 60. J, Navarro JL, San Miguel JF. Flow cytometric analysis of mast cells 35 Olweus J, BitMansour A, Warnke R, Thompson PA, Carballido J, from normal and pathological human bone marrow samples. Am Picker LJ, Lund-Johansen F. Dendritic cell ontogeny: a dendritic J Pathol 1996; 149: 1493–1499. cell lineage of myeloid origin. Proc Natl Acad Sci USA 1997; 94: 25 Li JL, Shen GL, Gethie MA, May RD, Till M, Gethie V, Uhr JW, 12551–12556. Janossy G, Thorpe PE, Amlot P, Vitetta ES. The epitope specificity 36 Gore SD, Kastan MB, Civin CI. Normal human bone marrow pre- and tissue reactivity of four murine monoclonal anti-CD22 anti- cursors that express terminal deoxynucleotidyl transferase include bodies. Cell Immunol 1989; 118: 85–99. T cell precursors and possible lymphoid stem cells. Blood 1991; 26 Campana D, Janossy G, Bofill M, Trejdosiewicz LK, Ma D, 77: 1681–1690. Hoffbrand AV, Mason DY, Lebacq AM, Forster HK. Human B cell 37 Farahat N, Lens D, Zomas A, Morilla R, Matutes E, Catovsky D. development. I. Phenotypic differences of B lymphocytes in the Quantitative flow cytometry can distinguish between normal and leukaemic B cell precursors. Br J Haematol 1995; 91: 640–646. bone marrow and peripheral lymphoid tissue. J Immunol 1985; 38 Galy A, Travis M, Cen D, Chen B. Human T, B, NK and dendritic 134: 1524–1530. cells arise from a common bone marrow progenitor cell subset. 27 Van Dongen JJM, Breit TM, Adriaansen HJ, Beishuizen A, Immunity 1995; 3: 459–473. Hooijkaas H. Detection of minimal residual disease in acute leu- 39 Pontvert-Delucq S, Breton-Gorius J, Schmitt C, Baillou C, Guich- kemia by immunological marker analysis and polymerase chain ard J, Najman A, Lemoine FM. Characterization and functional reaction. Leukemia 1992; 6 (Suppl. 1): 47–59. analysis of adult human bone marrow cell subsets in relation to 28 Campana D, Pui CH. Detection of minimal residual disease in B-lymphoid development. Blood 1993; 82: 417–429. acute leukemia: methodological advances and clinical signifi- 40 Hurwitz CA, Loken MR, Graham ML, Karp JE, Borowitz MJ, Pullen cance. Blood 1995; 85: 1416–1434. DJ, Civin CI. Asynchronous antigen expression in B lineage acute 29 Boue´ DR, LeBien TW. Expression and structure of CD22 in acute lymphoblastic leukemia. Blood 1988; 72: 299–307. leukemia. Blood 1988; 71: 1480–1486. 41 Drach J, Drach D, Glassl H, Gattringer C, Huber H. Flow cyto- 30 Uckun FM, Ledbetter JA. Immunobiologic differences between metric determination of atypical antigen expression in acute leu- normal and leukemic human B cell precursors. Proc Natl Acad kemia for the study of minimal residual disease. Cytometry 1992; Sci USA 1988; 85: 8603–8607. 13: 893–901. 31 Nuss R, Kitchingman G, Cross A, Zipf TF, Antoun GR, Bernstein 42 Orfao A, Ciudad J, Lopez-Berges MC, Lopez A, Vidriales B, I, Behm F, Pullen DJ, Crist W, Mirro J Jr, Goorha R. T cell receptor Caballero MD, Valverde B, Gonzalez M, San Miguel JF. Acute gene rearrangement in B precursor acute lymphoblastic leukemia lymphoblastic leukemia (ALL): detection of minimal residual dis- correlate with age and stage of B cell differentiation. Leukemia ease (MRD) by flow cytometry. Leuk Lymphoma 1994; 13: 87–90. 1988; 2: 727. 43 Van Wering ER, Beishuizen A, Roeffen ETJM, van der Linden- 32 Comans-Bitter WM, Versteeg S, Hooijkaas H, Van Dongen JJM. Schrever BEM, Verhoeven M-AJ, Hahlen K, Hooijkaas H, van CD22 expression in precursor B cells tested with the CD22 panel Dongen JJM. Immunophenotypic changes between diagnosis and antibodies. In: Schlossman S et al (eds). Leucocyte Typing V. relapse in childhood acute lymphoblastic leukemia. Leukemia Oxford University Press: Oxford, 1995, pp 528–530. 1995; 9: 1523–1533.