Multiparameter Phenotype Mapping of Normal and Post-Chemotherapy B

Multiparameter phenotype mapping of normal and post-chemotherapy B lymphopoiesis in pediatric bone marrow MN Dworzak, G Fritsch, C Fleischer, D Printz, G Fro¨schl, P Buchinger, G Mann and H Gadner

Children’s Cancer Research Institute, St Anna Kinderspital, Kinderspitalgasse 6, A-1090 Vienna, Austria

We studied the differentiation profiles of B cell precursors Materials and methods (BCP) in normal and post-chemotherapy pediatric bone marrow (BM) using multiparameter flow cytometry. The goal of our study was to draw a comprehensive phenotypic map of the Sample description three major maturational BCP stages in BM. By correlating lineage-associated markers, CD45RA, and several adhesion mol- Pediatric bone marrow (BM) samples (n = 44) were obtained ecules, the stage-specific patterns were found to differ in cer- from children undergoing diagnostic BM aspiration for the fol- tain details from previously published concepts. Among the earliest BCP, a subset of CD34+CD10lo precursors was repeat- lowing diagnoses in the absence of BM involvement at the edly observed in addition to the well characterized time of immunologic investigation: suspicion of storage dis- CD34+CD10hiCD19+ majority of cells. Only two-thirds of these ease (n = 1), bacterial lymphadenitis (n = 1), neuroblastoma CD34+CD10lo cells expressed CD19. However, uniformity of (n = 1), severe aplastic anemia in remission (n = 2), single sys- phenotypic features, absence of T lineage markers, and the tem Langerhans cell histiocytosis (n = 1), idiopathic throm- regeneration kinetics after chemotherapy suggest the B lineage = + lo bocytopenic purpura (n 1), early stage (localized) non- affiliation of the CD34 CD10 precursors in general. In the more = = mature BCP, expression of CD10, CD20, cytoplasmic and sur- Hodgkin lymphoma (T-NHL: n 2, B-NHL: n 1) as well as face ␮ chains (c␮ and s␮) was observed to overlap more than acute leukemia in follow-up with confirmed complete clinical previously recognized. We found that CD20 and c␮ appear remission at the time of immunologic investigation (B lineage early during B cell ontogeny (already on CD34+ BCP), and that ALL: n = 32, T-ALL: n = 1, AML: n = 1). The median age of the CD10 is lost late, following the onset of s␮ expression. Differ- 2 patients at BM aspiration was 512 years (range: 5 months to ences between normal and post-chemotherapy BM specimens 7 1612 years). Among the leukemia/lymphoma patients after regarding the phenotypic appearance of BCP were exclusively = due to differences in the subset composition, as post-chemo- chemotherapy (n 35), the median time from remission to BM 3 therapy samples showed a preponderance of immature stages. sampling was 11 months (range: 2 months to 412 years). The Our observations may build a framework for comparing leu- five patients with the shortest time from remission to BM aspir- kemic cells with their normal counterparts to define possible ation (1 or 2 months) have subsequently stayed event free until leukemia-associated aberrations useful for residual disease the last evaluation, 10 months to 1 7 years later. Only two studies. 12 Keywords: B cell precursor; lymphopoiesis; bone marrow; pheno- patients have relapsed after the immunologic evaluation type; flow cytometry (AML) M2 1 month later, C-ALL 4 months later). The median observation time from BM aspiration to final evaluation was 2 6 112 years (range: 1 month to 212 years). The two patients with the shortest follow-up (1 and 2 months) have been in con- Introduction 3 firmed remission for 312 years and 4 months, respectively. All leukemia/lymphoma patients described in this paper were The mainstays in defining the phenotypic patterns of B cell treated according to BFM (Berlin–Frankfurt–Munster) treat- differentiation have already been documented during the last ¨ ment protocols. In brief, for ALL a multi-agent induction ther- two decades.1–4 In brief, the CD19 antigen is expressed apy administered in three cycles over the first six months was throughout all stages of B cell maturation,1,5 while CD10 is highly expressed in the most immature (CD34+) B cell precur- followed by a low-dose maintenance therapy until completion sors (BCP),3 and is lost at a later stage concomitantly with gain of the second therapy year. All data concerning these patients of surface expression of CD20, IgM and CD22.2 More were obtained from the Austrian study center of the recently, this classical model of B cell differentiation was chal- international BFM study group. lenged, as new antigenic patterns,6–8 new insights into the sequence of antigen acquisition in B cell ontogeny,9 and tissue-related differences in the coordinate patterns of B Antibodies lymphoid antigen expression were documented.10,11 In the present study, our aim was comprehensively to rede- Fluorescein isothiocyanate (FITC)-, and phycoerythrin (PE)- fine the phenotypic patterns of the early stages of normal B labeled, or unconjugated pure monoclonal antibodies cell differentiation in postnatal bone marrow by multi- (MoAbs) were used: CD7 (DK24-FITC), CD10 (SS2/36-PE), parameter flow cytometry. By comparing normal and post- CD11a (MHM24 pure), CD19 (HD37-FITC, pure), CD20 (B- chemotherapy bone marrow, we focused on sample-inherent Ly1-FITC), CD44 (DF1485 pure) and CD45 (T29/33-FITC) differences in BCP phenotype and subset composition. Our were all from Dako (Glostrup, Denmark); CD34 (HPCA2- approach could be of particular interest for the development FITC, -PE) fom Becton Dickinson (BD; Sunnyvale, CA, USA); of techniques for immunologic residual disease investigations CD7 (3A1-RD1-PE) and CD45RA (2H4-RD1-PE, pure) from in B lineage acute lymphoblastic leukemia (ALL). Coulter Immunology (Krefeld, Germany); CD49d (HP2/1 pure), and anti-human TdT (H Tdt-1, 3, 4-FITC) from Immuno- tech (Marseille, France); CD49e (P1D6 pure) from Oncogene Correspondence: MN Dworzak Sciences (Cambridge, MA, USA); CD99 (culture supernatant Received 20 September 1996; accepted 10 April 1997 of the hybrid cell line 12E7) from R Levy (Stanford University, Phenotypic map of B lymphopoiesis in pediatric BM MN Dworzak et al

Ј 1267 CA, USA). FITC-conjugated F(ab )2 fragments of polyclonal Results rabbit anti-human ␮-heavy-chain antibodies were from Dako. Deﬁnition and quantiﬁcation of BCP stages

Immunofluorescence staining procedure The BCP content of the BM specimens was estimated by staining with CD19. A median of 16% of MNC was CD19+ BCP Mononuclear cells (MNC) were isolated from the BM speci- = mens by density gradient centrifugation and washed twice in (range 2.9–47.6%, n 32). For a further subdivision, CD19 RPMI 1640 medium containing 2% fetal calf serum (FCS; both was correlated with CD45RA, which has been shown from Gibco, Paisley, UK) prior to immunofluorescence labe- (together with CD45) to define three major stages of BCP due 12 ling. Staining protocols employed three-color investigations in to differences in antigen density. CD99 was added especially for the definition of the most immature BCP (stage all experiments. All antibodies were used at concentrations + +++ × 5 1, CD45RA CD99 ; Figure 1, black) as reported pre- titrated for optimal staining. In brief, 1 to 2 10 MNC in 13,14 conventional analyses as well as at least 5 × 105 cells per viously. The next developmental stage (BCP 2, blue) was CD45RA++CD99+/++, and the most mature stage (BCP 3, red) analysis in gating experiments, were first incubated with the +++ + ++ Ј CD45RA CD99( )/ (see Figure 1). According to this subdiv- unconjugated MoAbs and then with biotinylated rabbit F(ab )2 anti-mouse Ig antibodies (from Dako). To exclude non-specific ision, any difference in the marker profiles of BCP from differ- staining via unsaturated binding sites of the second step ent BM specimens could be attributed exclusively to differ- reagent, this was followed by incubation with unlabeled irrel- ences in the subset composition. We found that specimens of ALL patients in remission and regenerating after induction evant murine IgG1 (MOPC21; Sigma, St Louis, MO, USA). Without washing, the directly labeled CD MoAbs were then therapy, were reduced in their total B cell content, but had a added simultaneously and together with streptavidin-R670 considerable relative expansion of stage 1 BCP (Table 1). (Gibco). Murine irrelevant isotype-matched fluorescent Patients at the end of low-dose maintenance chemotherapy as Ј well as the non-leukemic probands (without any prior MoAbs as well as irrelevant F(ab )2 fragments of rabbit Ig-FITC (all from Dako) were always included in the staining protocols chemotherapy) showed a preponderance of more mature as negative controls. Incubations were performed at 4°Cin stages. Due to the relatively small number of normal BM the dark over 30 min and were followed by washing steps donors, age-related differences were not assessed. using PBS (0.13 mol/l NaCl, 7 mmol/l Na2HPO4, 3 mmol/l NaH2PO4) with 2% FCS. Cellular permeabilization for the investigation of TdT and cytoplasmic ␮-chain expression was Phenotypic analysis of stage 1 BCP cells performed using a commercially available formaldehyde- 3 based erythrocyte lysing solution (BD) supplemented with Tri- The earliest B cells in the BM express CD10 and CD34. In ton-X (0.015%). The concentration of the latter had been accordance, these markers were found on the corresponding titrated for optimal permeabilization together with a good BCP subsets in our analysis (stage 1). A majority of stage 1 preservation of scattering and fluorescence properties. The cells typically displayed high levels of CD10 positivity (stage permeabilization step was always done after labeling of sur- 1B; Figure 2, green subset). However, precursor cells (11.5% = face antigens. Prior to flow cytometry, cell suspensions were median of total BCP 1; range 0–41%, n 19) with consider- passed through a 30 ␮m mylon mesh (Swiss Silk Bolting Mfg, ably lower expression of CD10 were also observed (Figure 2, lo + Zurich, Switzerland). black subset). These CD10 CD34 cells (stage 1A) expressed CD19 heterogeneously: a median of 39% was negative (stage 1A/0; range 0–91%). Irrespective of CD19 expression, how- Flow cytometry ever, all the CD10loCD34+ BCP expressed CD99 at the highest levels of all BM cells, and were positive with TdT (not shown). Analyses were performed on a FACStar Plus (BD) equipped Compared to BCP 1B, they appeared to stain slightly more with an Argon laser tuned to 488 nm. Calibration beads intensely for CD99, and to display somewhat larger FSC as (Calibrite; BD) were routinely used for monitoring and optim- well as SSC properties (not shown). The CD10loCD34+ cells izing the instrument settings. Laser as well as photomultiplier were more numerous in samples with an expanded total BCP parameters were kept constant for all experiments. Data were 1 compartment, and the CD19− stage 1A/0 cells were more acquired with the LYSYS II or CELL Quest software (BD). For- numerous in parallel with an expanded total stage 1A ward light scattering (FCS), orthogonal light scattering (SSC), (P = 0.02), illustrating parallel patterns of generation kinetics and fluorescence signals (FL-1-FITC, FL-2-PE, FL-3-Red670) for all these putative BCP subsets. were stored in listmode data files. Each conventional measure- CD7 has been shown to be expressed on rare subsets of ment contained 30 000 cells. For better definition of small CD34+CD10+ lymphoid precursors in adult BM.15 In conven- subsets of cells, gated acquisition was performed for some tional experiments (n = 11 samples), no CD7+CD34+ double- marker combinations. Samples were first analyzed conven- positive cells could be observed. In gating experiments (n = 5 tionally, and using the same material, life gates were then set samples; correlating the markers CD34, CD7, CD19, CD10 in the appropriate fluorescence vs SSC correlations and gated and CD99), no CD7 reactivity reliably above background was events acquired to maximum yield (at least 10 000 events). All found on CD34+CD10+ precursors in general (data not data were analyzed using the PAINT-A-GATE or the CELL shown). Quest software (BD). CD20 was expressed at low or intermediate density on a considerable proportion of stage 1 cells (31.5% median; range Statistical analysis 7–65%; Figure 3a; black subset). It is noteworthy that positivity with CD20 could be demonstrated on cells of both stage Correlations concerning the size of different subsets of BCP 1 subsets (A and B). In staining for surface or cytoplasmic ␮- cells were estimated using the Spearman’s rank correlation heavy chains (s␮ or c␮), no reactivity was seen in conven- coefficient test. tional (ungated) analyses. Since the combined expression of Phenotypic map of B lymphopoiesis in pediatric BM MN Dworzak et al 1268

Figure 1 General stage definition of B cell precursors in BM by three-color flow cytometric analysis. MNC obtained by density gradient separation were stained with MoAbs CD19, CD45RA and CD99. Thirty thousand cells were acquired on a FACStar Plus and analyzed with the CELL Quest software. Data are presented as dotplots of log fluorescence. B cells (CD19+) at various stages of maturation are depicted in colors, whereas all non-B lineage cells appear grey. Three major subsets of BCP were defined according to their differential expression of CD45RA and CD99: stage 1 (most immature; black), stage 2 (blue), and the most mature subset (stage 3, red).

Figure 2 Three-color ﬂow cytometric analysis of BCP stage 1. MNC were stained in various combinations with CD10, CD19, CD34 and CD99. Thirty thousand cells were acquired. Only CD34+CD10+ stage 1 BCP are displayed painted (green and black), whereas all other cells appear grey. CD10+CD34+ cells could be subdivided according to levels of CD10 expression: CD10loCD34+ cells (stage 1A; black subset) show very strong expression of CD99, but CD19 negativity in a subpopulation (stage 1A/0). Note that the sample displayed is extraordinarily rich in these cells. The CD19hiCD34+ cells (stage 1B; green) are positive with CD19 throughout, and represent the major immature BCP population extensively characterized in the classical studies of B lymphopoiesis.

CD34 and c␮ seems to be a very rare event in normal B cell parallel. A median of 5.3% of BCP 1 cells expressed ␮-chains differentiation,16 we performed gated acquisition experiments in their cytoplasm (range 1.8–6.4%), whereas no reactivity on ﬁve samples. At least 10000 CD34+ cells, counterstained was detected on the cell surface (Table 2). The highest pro- with CD19 and anti-␮-heavy chain antibodies with (for c␮)or portions of CD34+c␮+ cells were found in those two samples without (for s␮) prior permeabilization, were selectively with the highest proportion of BCP 1 cells relative to total B acquired. Appropriate negative controls were always run in cells. Both samples were from patients in profound hemato- Phenotypic map of B lymphopoiesis in pediatric BM MN Dworzak et al 1269 Table 1 Comparison of the BCP subset composition in the BM of ALL patients after hematopoietic regeneration from high-dose induction treatment as well as after low-dose maintenance therapy, and of probands without prior chemotherapy or bone marrow disease

BCP/MNC% % BCP 1 % BCP 2 % BCP 3 median (range) median (range) median (range) median (range)

After induction 11 (3–31) 59 (33–100) 26 (0–59) 8 (0–31) n = 9 After maintenance 18 (3–37) 18 (4–36) 59 (44–94) 20 (0–47) n = 8 Without chemotherapy 16 (5–40) 9 (9–21) 40 (14–63) 41 (28–77) n = 6

Three CD19+ BCP subsets at different stages of maturity were distinguished by differential expression of CD45RA and CD99. Note that patients after induction therapy have a diminished proportion of total BCP relative to total mononuclear bone marrow cells, but show a preponderance of immature BCP cells reflecting B lymphopoietic regenerative activity. poietic regeneration after previous chemotherapy for c␮− terns. CD49e was found to be expressed especially during leukemia. stage 1. The antigen appeared to be downregulated to low levels on BCP 2 cells and was absent from the surface of stage 3 cells. Stage 3 cells displayed by far the most intense CD44 Phenotypic analysis of stage 2 and stage 3 BCP cells expression, and stage 1 BCP by far the highest CD99 density, while stage 2 cells were in part lowest in expression of these The two more mature stages of BCP cells were CD34−. Stage two antigens. 2 cells (Figure 3; blue subset) showed generally lower CD10 expression compared to the CD10hi BCP of stage 1 (Figure 3; black subset), and a variable CD20 antigen density ranging Discussion from negative to intermediate levels. In contrast, homo- geneous positivity for c␮ was observed throughout this parti- Several concepts of the orderly expression of antigens during cular stage. Only a minority of stage 2 cells also presented ␮ normal human B cell precursor differentiation have been chains on their surface (12% median; range 0–26.5%, n = 10). presented during the last 15 years.1–4,10,11 In certain details, With transition from stage 2 to stage 3, further changes in substantial differences exist between these studies. Major top- expression of several antigens were observed. Cells of the lat- ics of discussion have been the sequence of acquisition of the ter, most mature stage (Figure 3A; red subset) displayed a gain earliest B lymphoid antigens CD10 and CD19,1,2,4,9,24 as well in expression of the antigens CD45RA and especially CD20, as the coordinated phenotypic changes concerning the loss of which were uniformly expressed at high density on all BCP CD10 expression in correlation with the sequence of acqui- 3. All stage 3 BCP were also s␮+. These phenotypic shifts were sition of the antigens CD20 and IgM.2–4,10,11 Obviously, a sig- intimately associated with a further reduction in CD10 den- nificant deal of variation has been due to methodological sity. We observed a continuous gradient of expression of issues (ie normal or leukemic samples as reference, two- or expression of this antigen on stage 3 populations, suggestively three-color investigations, affinity of the MoAbs used) as well reflecting the final loss of CD10 expression during this stage as to tissue-related differences in the proportions of BCP (Figure 3b). A median of 45% (range 6–84.5%, n = 14) of stage subpopulations at different maturational stages. 3 cells were still CD10+ (violet subset in Figure 3b). The Our observations are pertinent to these issues. The first remaining stage 3 cells were negative for CD10 (red subset in major finding of our study is that normal and regenerating BM Figure 3b). Significantly (P Ͻ 0.02), CD10+ stage 3 BCP were samples apparently differ in the subset composition of their B more numerous in the samples with higher proportions of cell pool, with a preponderance of immature BCP in post- more immature BCP cells. Considering a presumed differen- chemotherapy specimens. There were, however, no differ- tiation sequence from CD10+ to CD10− in stage 3, this could ences between the two sample sources regarding stage-spe- be interpreted as a ‘shift to the left’ in BCP 3 due to the cific phenotypic profiles of BCP. A second major observation generation kinetics of the B cell pool in general. is the considerable heterogeneity of the earliest stage of BCP. The majority of these cells adhered to the well-characterized CD10hiCD34+CD19+ phenotype,3 while CD34+ cells with Stage-related differential expression of adhesion lower CD10 expression were also found. Unequivocal B lin- molecules on BCP subsets eage affiliation of the latter could be proved by virtue of CD19 expression in only two-thirds of this population. Importantly, We investigated the adhesion molecules CD11a, CD44, these cell types could be observed not only in BM specimens CD49d and CD49e, of which maturation-related differences from patients after chemotherapy, but also in normal controls. in expression as well as major roles in B cell development The data are in line with recent publications which showed and function have been reported,17–21 as well as CD99, a that the proportion of CD10+ cells exceed those of CD19+ novel CD antigen involved in T cell adhesion processes.22,23 cells in adult BM samples enriched for early precursors.15,25 Two general patterns of stage-related differential expression However, a lineage relationship of the majority of the were recognized (Figure 4). Roughly, the surface density of the CD10loCD34+ precursors, irrespective of CD19 expression, is antigens CD11a and CD44 increased with maturation, while now suggested according to their uniform phenotypic appear- that of CD49d, CD49e and CD99 decreased. Differences in ance concerning CD99, TdT, and light scattering properties, expression of CD11a as well as CD49d on the three major as well as their generation kinetics. In order to confirm the B BCP stages were minor with broadly overlapping density pat- lineage affiliation of the CD19−CD10loCD34+ precursors, we Phenotypic map of B lymphopoiesis in pediatric BM MN Dworzak et al 1270

Figure 3 Comprehensive three-color analysis of the phenotypic patterns of BCP differentiation. Expression patterns of several lineage-restricted and -associated antigens on cells of the three major BCP stages (stage 1: black, stage 2: blue and stage 3: red) are shown. All non-B lineage cells appear in grey. (a) MNC were stained with CD19 and CD45RA in combination with CD10, CD20, CD34, or MoAbs to ␮ chains (with (for c␮) and without (for s␮) prior permeabilization). BCP stages were distinguished according to levels of CD45RA expression. Note considerable expression of CD20 on stage 1 cells. (b) MNC were stained with CD10 and CD45RA in combination with CD20 or MoAbs to ␮ chains (with and without prior permeabilization). Coexpression patterns of CD10 with CD20, c␮ and s␮ are shown. BCP stages are displayed in the same colors as in (a) except for stage 3: the cells of this stage could be subdivided according to CD10 expression (CD10dim = violet; CD10− = red). Note that in stage 2 (blue) all cells are c␮ + but in a subpopulation lack CD20 expression.

excluded CD7 expression and, hence, T lineage commitment, CD79a. This would support the opinion deduced from in vitro although rare CD7+ prothymocytes were recently tracked in differentiation studies with fetal liver B cells, that CD10 the TdT+ fraction of adult BM.6,15 Myeloid commitment of appears earlier than CD19 on the surface of B cell precur- CD10+ precursors, on the other hand, has been ruled out in sors.9,11 It would also refine the model of the earliest steps of the literature.26 Since CD79a seems to be of unique B cell B cell differentiation in the BM and suggest that very immature specificity,27 definite B lineage affiliation should be confirmed B lineage cells are CD34+CD10lo, then acquire CD19, and especially by testing for this antigen. Further investigations are subsequently develop into the CD34+CD10hiCD19+ pheno- under way, but we already have preliminary data which indi- type. This sequence of early antigen acquistion of BCP has cate that these CD19−CD10loCD34+ precursors do express been under repeated debate, and the simultaneous appear- Phenotypic map of B lymphopoiesis in pediatric BM MN Dworzak et al 1271

Figure 4 Three color analysis of stage-speciﬁc adhesion molecule expression during BCP maturation. MNC were stained with CD19 and CD45RA in combination with CD11a, CD44, CD49d, CD49e or CD99. The three major stages of BCP development (stage 1: black; stage 2: blue; stage 3: red) were addressed according to positivity with CD45RA. Non-B cells appear grey. Note incremental and decremental changes in expression of the investigated adhesion structure during BCP maturation.

Table 2 Analysis of cytoplasmic as well as surface ␮-heavy chain expression of stage 1 B cell precursors

Patient No. Diagnosis Age %BCP 1 %BCP 1 c␮-pos c␮-pos s␮-pos s␮-pos (years) of total of total of BCP 1 BCP 1 of of BCP 1 BCP 1 of MNC BCP (%) MNC (%) (%) MNC (%)

9 1 SAA 412 2.8 13.7 1.8 0.05 ND ND in remission 6 2 C-ALL 612 3.1 9.3 5.3 0.16 ND ND remission: 3 y 1 3 C-ALL 412 6.6 58.9 6.4 0.42 0 0 remission: 6 mo 4 C-ALL 5 17.9 59.5 5.8 1.04 0 0 remission: 4 mo 9 5 C-ALL 512 3.2 11.3 3.7 0.12 ND ND remission: 3 y

MNC of five patients in confirmed clinical remission from hematological disease at the time of phenotypic analysis were stained with MoAbs to CD19, CD34, and ␮ chains with or without prior permeabilization. In conventional as well as gating experiments (10 000 CD34+ cells were selectively acquired), the proportions of BCP 1 cells (CD19+CD34+), total B cells (CD19+), and of ␮ chain+ (cytoplasmic or surface) CD19+CD34+ BCP (relative to total MNC) were determined. The data show that CD19+CD34+ stage 1 cells are c␮+ in a considerable proportion of cells but lack s␮ expression entirely. Note that in all ALL follow-up patients, the leukemic cells at primary diagnosis were lacking c␮ expression except for proband No. 5 (3 years in remission; initially 5% of blasts c␮+). ance of CD10 and CD19,2,3,24 as well as the precedence of that CD10 is lost prior to c␮ expression.3 We could demon- CD19 have formerly been considered.1,4 strate that CD20 and c␮ already appear on a proportion of the A further major finding of our study is that CD10 broadly most immature BCP subset, which adds to the heterogeneity of overlaps in expression with CD20, c␮, and (to a lesser extent) stage 1. After transition to stage 2, which is still CD10+, essen- s␮. Our observations contrast with the Loken model2 of B cell tially all BCP cells were found to acquire c␮ positivity. CD20 differentiation which considers that CD20 appears very late expression was far more variable in this subset, ranging from on postnatal BCP together with s␮, and also with the notion negative to relatively high levels. Such a desynchronized regu- Phenotypic map of B lymphopoiesis in pediatric BM MN Dworzak et al 1272 lation of CD20 and c␮ appears as an example for the concept 3 Ryan D, Kossover S, Mitchell S, Frantz C, Hennessy L, Cohen H. of flexible developmental programs with multiple differen- Subpopulations of common acute lymphoblastic leukemia anti- tiation options in human B lymphopoiesis.11 Only a minority gen-positive lymphoid cells in normal bone marrow identified by ␮ hematopoietic differentiation antigens. Blood 1986; 68: 417–425. of stage 2 cells, on the other hand, displayed a s expression, 4 Hokland P, Ritz J, Schlossman SF, Nadler LM. Orderly expression suggesting a later emergence of this antigen during B cell of B cell antigens during the in vitro differentiation of nonmalig- development. With maturation to stage 3, however, s␮ nant human pre-B cells. J Immunol 1985; 135: 1746–1751. appeared on the vast majority of cells, coincidently with upre- 5 Nadler LM, Anderson KC, Marti G, Bates M, Park E, Daley JF, gulation of CD20 to very high levels and with a further down- Schlossman SF. B4, a human B lymphocyte-associated antigen regulation of CD10. Hence, we document that CD10 is still expressed on normal, mitogen-activated, and malignant B lymphocytes. J Immunol 1983; 131: 244–250. expressed on a subset of mature BCP, and that the changes of 6 Smith RG, Kitchens SR. Phenotypic heterogeneity of TDT+ cells in CD10 expression during B cell maturation are many-fold. the blood and bone marrow: implications for surveillance of Much of the interest in the phenotype of physiologic BCP is residual leukemia. Blood 1989; 74: 312–319. due to the efforts made to characterize phenotypic differences 7 Hurwitz CA, Gore SD, Stone KD, Civin CI. Flow cytometric detec- between them and their malignant counterparts, the blasts of tion of rare normal human marrow cells with immunophenotypes B lineage acute lymphoblastic leukemia (ALL).28–32 Such dif- characteristic of acute lymphoblastic leukemia cells. Leukemia ferences could be applicable in the immunologic investigation 1992; 6: 233–239. 8 Saeland S, Duvert V, Pandrau D, Caux C, Durand I, Wrighton N, of minimal residual disease (MRD), and could aid in the differ- Wideman J, Lee F, Banchereau J. Interleukin-7 induces the pro- 33 ential diagnosis between hyper-regeneration and relapse in liferation of normal human B-cell precursors. Blood 1991; 78: patients with typical BM lymphocytosis after chemo- 2229–2238. therapy.30,34,35 Several antigen patterns have been considered 9 Uckun FM, Ledbetter JA. Immunobiologic differences between in the literature as aberrant and leukemia-associated (eg the normal and leukemic human B-cells precursors. Proc Natl Acad pairwise expression of CD34 or TdT with CD20, CD22 or Sci USA 1988; 85: 8603–8607. ␮ 36,37 10 LeBien TW, Wo¨rmann B, Villablanca JG, Law C-L, Steinberg LM, c ). However, more recent work identified rare normal Shah VO, Loken MR. Multiparameter flow cytometric analysis of populations with such an antigen profile, qualifying them as human fetal bone marrow B cells. Leukemia 1990; 4: 354–358. unusual rather than aberrant.6–8,16,28,38 We also found coex- 11 Uckun FM. Regulation of human B-cell ontogeny. Blood 1990; pression of CD20 as well as of c␮ on a significant proportion 76: 1908–1923. of very immature BCP. As has been postulated,37 we now 12 Shah VO, Civin CI, Loken MR. Flow cytometric analysis of human document that these phenotypes can be readily discovered bone marrow. IV. Differential quantitative expression of T-200 in particular in post-chemotherapy regenerating pediatric BM common leukocyte antigen during normal hemopoiesis. J Immunol 1988; 140: 1861–1867. samples, which hampers an application in MRD investi- 13 Bodger MP, Francis GE, Delia D, Granger SM, Janossy G. A mono- gations. At least, the existence of such antigenic combinations clonal antibody specific for immature human hemopoietic cells as physiologic patterns of early B lineage differentiation pro- and T lineage cells. J Immunol 1981; 127: 2269–2274. motes our understanding of leukemia phenotypes as replicas 14 Dworzak MN, Fritsch G, Buchinger P, Fleischer C, Printz D, of normal BCP. It may explain the lack of an inverse corre- Zellner A, Scho¨ llhammer A, Steiner G, Ambros PF, Gadner H. lation between CD34 and CD20 in B lineage ALL36 and the Flow cytometric assessment of human MIC2 expression in bone 39 marrow, thymus, and peripheral blood. Blood 1994; 83: 415–425. high incidence of CD34 coexpression in the pre-B subtype. 15 Gore SD, Kastan MB, Civin CI. Normal human bone marrow pre- The investigation of multiple phenotypes and of differences cursors that express terminal deoxynucleotidyl transferase include in antigen density may help in developing new strategies for T-cell precursors and possible lymphoid stem cells. Blood 1991; distinguishing leukemic from normal cells.16,40–43 In this 77: 1681–1690. regard, our study portrays a comprehensive phenotypic map 16 Campana D, Pui C-H. Detection of minimal residual disease in of the early stages of normal B cell differentiation which may acute leukemia: methodologic advances and clinical significance. build a framework for comparative investigations in order to Blood 1995; 85: 1416–1434. 17 Kansas GS, Dailey MO. Expression of adhesion structures during highlight phenotypic aberrations of leukemic blasts to be used B cell development in man. J Immunol 1989; 142: 3058–3062. in MRD studies. 18 Ryan DH, Nuccie BL, Abboud CN, Winslow JM. Vascular cell adhesion molecule-1 and the integrin VLA-4 mediate adhesion of human B cell precursors to cultured bone marrow adherent cells. Acknowledgements J Clin Invest 1991; 88: 995–1004. 19 Miyake K, Medina KL, Hayashi S-I, Ono S, Hamaoka T, Kincade PW. Monoclonal antibodies to Pgp-1/CD44 block lympho-hemo- We are grateful to the hematooncology, laboratory and anes- poiesis in long-term bone marrow cultures. J Exp Med 1990; 171: thesia teams of the St Anna Kinderspital for their support and 477–488. for providing bone marrow specimens. Our special thanks, in 20 Ryan DH, Nuccie BL, Abboud CN. Inhibition of human bone mar- this regard, go to S Juhasz, R Kornmu¨ller and U Stalze. We row lymphoid progenitor colonies by antibodies to VLA integrins. are also indebted to J Regelsberger for providing the clinical J Immunol 1992; 149: 3759–3764. data, and to N Jones for critically reading the manuscript. 21 Hynes RO. Integrins: a family of cell surface receptors. Cell 1987; 48: 549–554. 22 Bernard A, Aubrit F, Raynal B, Pham D, Boumsell L. A T cell surface molecule different from CD2 is involved in spontaneous References rosette formation with erythrocytes. J Immunol 1988; 140: 1802–1807. 1 Nadler LM, Korsmeyer SJ, Anderson KC, Boyd AW, Slaughenhoupt 23 Bernard G, Zoccola D, Deckert M, Breittmayer J-P, Aussel C, B, Park E, Jensen J, Coral F, Mayer RJ, Sallan SE, Ritz J, Schlossman Bernard A. The E2 molecule (CD99) specifically triggers homo- SF. B cell origin of non-T cell acute lymphoblastic leukemia – typic aggregation of CD4+ CD8+ thymocytes. J Immunol 1995; a model for discrete stages of neoplastic and normal pre-B cell 154: 26–32. differentiation. J Clin Invest 1984; 74: 332–340. 24 Schmitt C, Eaves CJ, Lansdorp PM. Expression of CD34 on human 2 Loken MR, Shah VO, Dattilio KL, Civin CI. Flow cytometric analy- B cell precursors. Clin Exp Immunol 1991; 85: 168–173. sis of human bone marrow. II. Normal B lymphocyte develop- 25 Saeland S, Duvert V, Caux C, Pandrau D, Favre C, Valle A, Dur- ment. Blood 1987; 70: 1316–1324. and I, Charbord P, de Vries J, Banchereau J. Distribution of sur- Phenotypic map of B lymphopoiesis in pediatric BM MN Dworzak et al 1273 face-membrane molecules on bone marrow and cord blood Loonen AH, Broekema GJ, de Waal FC, Veerman AJP. Immuno- CD34+ hematopoietic cells. Exp Hematol 1992; 20: 24–33. logical phenotype of lymphoid cells in regenerating bone marrow 26 LeBien TW, McCormack RT. The common acute lymphoblastic of children after treatment for acute lymphoblastic leukemia. Eur leukemia antigen (CD10) – emancipation from a functional J Haematol 1988; 41: 170–175. enigma. Blood 1989; 73: 625–635. 36 Ryan DH, Chapple CW, Kossover SA, Sandberg AA, Cohen HJ. 27 Mason DY, Cordell JL, Tse AGD, van Dongen JJM, van Noesel Phenotypic similarities and differences between CALLA-positive CJM, Micklem K, Pulford KAF, Valensi F, Comans-Bitter WM, acute lymphoblastic leukemia cells and normal marrow CALLA- Borst J, Gatter KC. The IgM-associated protein mb-1 as a marker of positive B cell precursors. Blood 1987; 70: 814–821. normal and neoplastic B cells. J Immunol 1991; 147: 2474–2482. 37 Hurwitz CA, Loken MR, Graham ML, Karp JE, Borowitz MJ, Pullen 28 Janossy G, Bollum FJ, Bradstock KF, McMichael A, Rapson N, DJ, Civin CI. Asynchronous antigen expression in B lineage acute Greaves MF. Terminal transferase-positive human bone marrow lymphoblastic leukemia. Blood 1988; 72: 299–307. cells exhibit the antigenic phenotype of common acute lymphobl- 38 Campana D, Coustan-Smith E, Janossy G. The immunologic detec- astic leukemia. J Immunol 1979; 123: 1525–1529. tion of minimal residual disease in acute leukemia. Blood 1990; 29 Janossy G, Bollum FJ, Bradstock KF, Ashley J. Cellular phenotypes 76: 163–171. of normal and leukemic hemopoietic cells determined by analysis 39 Pui C-H, Hancock ML, Head DR, Rivera GK, Look T, Sandlund with selected antibody combinations. Blood 1980; 56: 430–441. JT, Behm FG. Clinical significance of CD34 expression in child- 30 Greaves MF, Hariri G, Newman RA, Sutherland DR, Ritter MA, hood acute lymphoblastic leukemia. Blood 1993; 82: 889–894. Ritz J. Selective expression of the common acute lymphoblastic 40 Van Wering ER, Beishuizen A, Roeffen ETJM, van der Linden- Schrever BEM, Verhoeven M-AJ, Hahlen K, Hooijkaas H, van leukemia (gp100) antigen on immature lymphoid cells and their ¨ Dongen JJM. Immunophenotypic changes between diagnosis and malignant counterparts. Blood 1983; 61: 628–639. relapse in childhood acute lymphoblastic leukemia. Leukemia 31 Greaves MF. Differentiation-linked leukemogenesis in lympho- 1995; 9: 1523–1533. cytes. Science 1986; 234: 697–704. 41 Lavabre-Bertrand T, Janossy G, Ivory K, Peters R, Secker-Walker 32 Campana D, Janossy G, Bofill M, Trejdosiewicz LK, Ma D, L, Porwit-MacDonald A. Leukemia-associated changes identified Hoffbrand AV, Mason DY, Lebacq A-M, Forster HK. Human B cell by quantitative flow cytometry: I. CD10 expression. Cytometry development. Phenotypic differences of B lymphocytes in the 1994; 18: 209–217. bone marrow and peripheral lymphoid tissue. J Immunol 1985; 42 Kreindler D, Petsche D, Hrincu A, Gougos A, Quackenbush EJ, 134: 1524–1530. Freedman MH, Gelfand EW, Letarte M. Quantitative phenotyping 33 Campana D, Coustan-Smith E, Behm FG. The definition of of childhood leukemia identifies variable and invariable cell sur- remission in acute leukemia with immunologic techniques. Bone face antigens. Leuk Lymphoma 1990; 3: 7–18. Marrow Transplant 1991; 8: 429–437. 43 Lavabre-Bertrand T, Duperray C, Brunet C, Poncelet P, Exbrayat 34 Pearl ER. Pre-B-cells in normal human bone marrow and in bone C, Bourquard P, Lavabre-Bertrand C, Brochier J, Navarro M, marrow from patients with leukemia in remission: persistent quan- Janossy G. Quantification of CD24 and CD45 antigens in parallel titative differences and possible expression of cell surface IgM in allows a precise determination of B-cell maturation stages: rel- vitro. Blood 1983; 61: 464–468. evance for the study of B-cell neoplasias. Leukemia 1994; 8: 35 Van den Doel LJ, Pieters R, Huismans DR, van Zantwijk CH, 402–408.