CD10 and CD19 fluorescence intensity of B-cell precursors in normal and leukemic bone marrow. Clinical characterization of CD10+strong and CD10+weak common acute lymphoblastic leukemia

Leukemia Research 23 (1999) 441 – 450

CD10 and CD19 fluorescence intensity of B-cell precursors in normal and leukemic bone marrow. Clinical characterization of CD10 + strong and CD10 + weak common acute lymphoblastic leukemia Eduardo Magalha˜es Rego a, Luiz Gonzaga Tone b, Aglair Bergamo Garcia a, Roberto Passetto Falca˜o a,* a

Department of Internal Medicine, Medical School of Ribeira˜o Preto, Uni6ersity of Sa˜o Paulo, CEP: 14049 -900, Ribeira˜o Preto SP, Brazil b Department of Pediatrics, Medical School of Ribeira˜o Preto, Uni6ersity of Sa˜o Paulo, CEP: 14049 -900, Ribeira˜o Preto SP, Brazil Received 29 June 1998; accepted 17 November 1998

Abstract In order to assess the age-related changes in CD10 and CD19 fluorescence intensity (FI) the present study analyzed by flow cytometry 56 sternal biopsies from ‘normal’ infants, children and adults undergoing cardiac surgery. The CD10 + weak subset was predominant in all age groups, representing approximately 50% of the bone marrow (BM) lymphoid cells in children younger than 4 years. Both CD10 + subsets significantly decreased with age but their ratio did not differ significantly. Moreover, the intensity of CD10 and CD19 fluorescence in the strong and weak subsets did not vary with age. The CD19 intensity was significantly higher in CD10 + weak than in CD10 + strong cells. In addition, we classified as CD10 + weak or CD10 + strong the leukemic cells from BM aspirates of 117 patients with common acute lymphoblastic leukemia (cALL) (78 children and 39 adults). A higher frequency of cases expressing the CD19 + CD10 + strong phenotype was observed both in children and adults. Children of the CD10 + weak group tended to be older than those of the CD10 + strong group (median =7 vs. 4 years, P= 0.07), and presented a significantly higher frequency of splenomegaly (93.7 vs. 55%, P =0.04), which was massive in about 60% of these cases. Among adults, a significantly higher frequency of cases expressing the CD10 + weak phenotype was observed in females. No other clinical or biological difference was detected between the two groups either for children or adults. Concerning the treatment outcome, we did not observe significant differences in complete remission rate (CRR) or in disease free survival (DFS) among the 32 children and 28 adults analyzed. Finally, we compared the CD10 and CD19 intensity in normal and leukemic BM. Overexpression of either or both antigens in leukemic cells was observed in 42.4% of the cALL cases. In these cases, using cut off values of 110 afu for the CD10 FI and of 100 afu for the CD19 FI, the detection of leukemic cells was possible at levels of 0.2% based on CD10 analysis, of 0.6% based on CD19, and 0.02% when both antigens were overexpressed. In conclusion, we demonstrated that the heterogeneity of CD10 and CD19 fluorescence intensity is of no clinical relevance in cALL, although its study may be helpful for the diagnosis and the detection of minimal residual disease. © 1999 Published by Elsevier Science Ltd. All rights reserved. Keywords: CD10; CD19; Acute lymphoblastic leukemia; Flow cytometry; Normal bone marrow; Fluorescence intensity; Minimal residual disease

1. Introduction Abbre6iations: ALL, acute lymphoblastic leukemia; cALL, common subtype of acute lymphoblastic leukemia; BM, bone marrow; CD, cluster of differentiation; FI, fluorescence intensity; CRR, complete remission rate; DFS, disease-free survival. * Corresponding author. Present address: Haematology Laboratory, University Hospital, Medical School of Ribeira˜o Preto, University of Sa˜o Paulo, CEP: 14049-900, Ribeira˜o Preto, SP, Brazil. Tel.: +55-16-6330436; fax: +55-16-6331144. E-mail address: [email protected] (R.P. Falca˜o)

During normal B-cell development, the intensity of CD10 expression decreases with maturation, with the more immature CD34 + precursors presenting higher densities of CD10 antigen [1,2]. Amongst leukemic lymphoblasts, heterogeneity in CD10 intensity has also been observed, which was associated with differences in immunophenotype, proliferative indexes and cytoge-

0145-2126/99/$ - see front matter © 1999 Published by Elsevier Science Ltd. All rights reserved. PII: S 0 1 4 5 - 2 1 2 6 ( 9 8 ) 0 0 1 9 0 - 8

442

E.M. Rego et al. / Leukemia Research 23 (1999) 441–450

netic abnormalities between CD10 + strong and CD10 + weak cells [3 – 7]. However, the clinical significance of this heterogeneity in common acute lymphoblastic leukemia (cALL) has not been fully characterized. Concerning the association between cytogenetic abnormalities and CD10 intensity, LavabreBertrand et al. [7] demonstrated that high levels were more frequent among hyperdiploid cases, low levels were associated with t(1;19), and undetectable CD10 was associated with t(4;11). Moreover, Glencross et al. [8] studied the correlation between sex and race distribution and CD10 density in leukemic cells from children with acute lymphoblastic leukemia (ALL) and demonstrated a higher frequency of cases expressing lower CD10 densities among Black boys when compared to White girls, whereas leukemic cells from Black girls and White boys expressed intermediate values. In addition, Look et al. [9] demonstrated a poorer prognosis for those patients whose leukemic lymphoblasts expressed lower CD10 levels during the S phase of the cell cycle. Yet this study analyzed only 35 children with cALL and a later study from the same Institution did not detect the expression of CD10 as an independent prognostic feature for children with B-lineage ALL [10]. Whether or not differences in CD10 intensity are of prognostic significance in cALL has not been definitely established, since most studies concerning the prognostic effect of CD10 expression compared the outcome of CD10 + versus CD10 − cases [10 – 18]. Nevertheless, more recently it has been shown that comparison of CD10 intensity between leukemic and normal B-cell precursors may be helpful for the immunological detection of minimal residual disease (MRD) [6,7]. In this way, Lavabre-Bertrand et al. [7], using a method of quantitative flow cytometry (quantimetry), demonstrated that the maximum antigen expression in fetal liver and normal bone marrow (BM) cells was 5 ×104 CD10 molecules/cell. Conversely, in about one third of B-lineage ALL cases, the median CD10 expression was above this value. In addition, Farahat et al. [6] quantified the number of TdT, CD10 and CD19 molecules per cell by direct quantitative flow cytometry and demonstrated that BM lymphoid cells from normal adults displayed CD10 intensities below 5×104 antigen molecules/cell, whereas in B-ALL cases the leukemic blasts displayed a higher number of CD10 molecules. In addition, leukemic cells displayed lower levels of TdT and higher levels of CD19 than their normal counterparts. In the present study we determined the distribution of B-progenitors CD10 + strong and CD10 + weak in normal BM biopsies from infants, children and adults, and clinically characterized cALL cases according to CD10 intensity. Finally, considering the potential use of the analysis of CD10 and CD19 intensities for MRD detection, we compared the fluorescence intensity (FI) of

these markers in normal lymphoid progenitors with those obtained in leukemic cells from 117 patients with cALL.

2. Patients and controls

2.1. Normal subjects We studied BM fragments obtained from the sternum of 44 children aged 2 weeks to 15 years undergoing cardiac surgery for the correction of congenital cardiopathies, and 12 adults aged 15–64 years, half of whom were operated for valvular heart disease and the other half for ischemic heart disease. Only patients with normal hematological counts, who exhibited no evidence of infections, autoimmune disorders or renal or hepatic dysfunction were considered eligible. A total of 21 patients were taking digoxin and/or furosemide at the time of the study. The subjects were divided into four groups according to age: Group 1: age 51 year (median=6 months; range =2 weeks-1 year); Group 2: 4 years ] age\1 year (2 years; 1.1–3.9 years); Group 3: 15 years ]age\ 4 years (7.7 years; 4–12 years); Group 4: age\15 years (34.5 years; 16–64 years). All samples were collected after obtaining informed consent from the donors or the person legally responsible for them, and the investigation was approved by the Ethics Committee of the University Hospital of the Medical School of Ribeira˜o Preto, University of Sa˜o Paulo.

2.2. Patients We studied BM aspirates from 117 patients (78 children and 39 adults) with ALL of the common subtype, whose immunophenotypic study was performed by the Hematology Laboratory of the University Hospital of the Medical School of Ribeira˜o Preto, University of Sa˜o Paulo. All cases were diagnosed according to the FAB proposal [19,20], and the immunophenotypic subset [21] was determined using a large panel of monoclonal antibodies, as previously described [22]. In addition, the scoring system proposed by Buccheri et al. [23,24] was adopted for lineage assessment. Only cases of ALL of the common subtype (CD10 + CD19 + ) were eligible. Since the detection of cytoplasmic immunoglobulin (cIg) was not performed routinely, both pre-B and early-pre-B + cases were included in the analysis. The clinical and biological features of these 117 patients were studied based on the records obtained at our Institution or from the referring hematologist, whereas the complete remission rate (CRR) and the disease free survival (DFS) were calculated only for the 32 children and 28 adults who were treated and followed up at the University Hospital and submitted to

E.M. Rego et al. / Leukemia Research 23 (1999) 441–450

443

the protocols GBTLI-85 (children) or L17-M (adults). The detailed drug schedule of the GBTLI-85 protocol and the results obtained with larger series of patients were described by Brandalise et al. [25], whereas for the L17-M protocol this information was reported by Clarkson et al. [26,27]. The mean follow-up of these patients was of 45.8 months (range: 10 – 88.5 months).

The contamination of the lymphoid gate by nucleated erythroid precursors was assessed by testing each sample with an antiglycophorin C monoclonal antibody employing an indirect immunofluorescence assay as previously described [28]. In the negative controls the same techniques were used, but the first antibodies were replaced with mouse IgG1× IgG2a of irrelevant specificity.

3. Material and methods

3.4. Analysis by flow cytometry

3.1. Bone marrow biopsies Bone fragments of ‘normal’ subjects measuring approximately 1× 0.3 cm were taken from the sternum immediately after thoracotomy and preserved in phosphate buffered saline (PBS) with 5% fetal calf serum (FCS) and 100 U of heparin. A cell suspension was prepared by rinsing the fragments with PBS+5% FCS and by teasing them apart with 23-gauge needles. The suspension was decanted for 3 min and the supernatant was centrifuged at 300 ×g for 10 min, at room temperature. The pellet was resuspended, the concentration was adjusted to 5× 106 cells/ ml, and 100 ml of this cell suspension was added to each test tube.

3.2. Bone marrow aspirates Bone marrow aspirates from patients with ALL were preserved in PBS with 5% FCS and 100 U of heparin. The samples were centrifuged at 300× g for 10 min at room temperature. The pellet was resuspended, the concentration was adjusted to 5× 106 cells/ml, and 100 ml of this cell suspension was added to each test tube.

3.3. Immunofluorescence techniques In order to assess the distribution of the lymphoid subsets in normal BM, and for the immunophenotypic classification of ALL cases, all samples were tested using a direct immunofluorescence technique using the following pairs of monoclonal antibodies (mabs): CD3×CD10; CD3 ×CD19; CD3 × CD(16/56); CD19 × CD10; CD20 ×CD5, CD7 ×CD11b, CD13× HLA-Dr, CD14× CD15; CD33 × CD34, anti-kappa × anti-lambda and mouse IgG1 ×IgG2a of irrelevant specificity were used as negative controls. The antiglycophorin C and anti-TdT antibodies were studied by one-color analysis. All antibodies were purchased already conjugated with phycoerythrin (PE) or fluorescein (FITC) from Becton Dickinson (Mountain View, CA), except for unconjugated antiglycophorin C, which was a kind gift from Prof. David Mason, John Radcliffe Hospital, Oxford, UK.

All samples were analyzed using a FACScan (Becton Dickinson, Immunocytometry Systems, San Jose, CA) equipped with an Argon ion laser with a wavelength setting of 488 nm. Between 10 000 and 30 000 events were collected in the lymphocyte gate, and the percentage of CD19 + /CD10 + cells was calculated after the exclusion of the background and of antiglycophorin C positive cells.

3.5. Determination of CD10 and CD19 fluorescence intensity The mab concentration sufficient to reach the maximum binding capacity was first determined and then the appropriate volume was added to each test tube. In order to obtain consistent fluorescent results during the time of the study, the flow cytometer was adjusted daily after running the Calibrite Beads (Becton Dickinson, Moutain View, CA) and a sample of frozen leukemic cells of known CD10 and CD19 FI. For the determination of FI the cells were incubated with the following pairs of mabs: CD3-PE× CD10-FITC, CD3-PE× CD19-FITC, CD10-FITC× CD19-PE, CD19-FITC× CD10-PE. The mean channel for green fluorescence was determined for the normal T-cell population and for the leukemic B-precursors. The FI was measured with detectors and amplifiers set on a logarithmic scale and expressed as arbitrary fluorescence units (afu) calculated by the formula: FI =

mean flourescent channel of the sample ×100 mean flourescent channel of the normal T−cell population

An arbitrary value of 35 relative units was adopted to distinguish CD19 + CD10 + strong from CD19 + + weak CD10 cells in normal and leukemic BM. For those cases (n= 5) where the normal T-cell population represented less than 1% of the gated events, the FI was calculated by using as denominator the background fluorescence of non B-cells stained with CD10. Finally, in order to rule out the possibility that any variation in the CD3 expression could affect the FI determination, we compared the CD10 FI values obtained using either CD2 or CD3 as a T-marker in 10 cALL cases and 10 normal B.M. samples (different from those previously described).

E.M. Rego et al. / Leukemia Research 23 (1999) 441–450

444

Table 1 Percentage of CD19+, CD19+/CD10+weak and CD19+/CD10+strong cells in the lymphocyte gate of BM biopsies from normal infants, children and adults Subset

Group 1 (age51 year) (%)

Group 2 (1 yearBage54 year) (%)

Group 3 (4 yearBage515 year) Group 4 (age\15 (%) year) (%)

CD19+ CD10+weak CD10+strong CD10+weak/ CD10+strong ratio

66.5a (44.1–70.4) 53.8 (37.2–74.4) 2.2 (1.3–3.6) 25.9 (13.6–51.2)

67.8 (53.5–78.1) 54.7 (43.6–63.4) 3.1 (2.2–3.7) 18.2 (14.7–26.2)

52.5 (41.6–58.4) 39.4 (30.9–50) 3.6 (2–4.4) 12.6 (10.1–16.1)

33.6c (16.5–40.3) 11.1d (5.8–17.1) 0.8d (0.3–1.2) 13.2 (11–20.9)

Pb

P50.05 P50.05 P50.05 P =0.13

a

Results= median (25th–75th percentiles). Calculated by the Kruskal–Wallis Test. c Values obtained for Group 4 were significantly different from those obtained for Groups 1 and 2. Calculated by Dunn’s Multiple Comparisons method. d Values obtained for Group 4 were significantly different from those obtained for Groups 1, 2 and 3. Calculated by Dunn’s multiple comparisons method. b

3.6. Statistical analysis In order to determine if the percentage of each subset had a Gaussian distribution we applied the Kolmogorov Smirnov test [29]. Since the data were distributed in a non-Gaussian pattern, we expressed the percentages of positivity using the median and the 25th and 75th percentiles. Kruskal – Wallis one-way analysis of variance (ANOVA) followed by Dunn’s method of multiple comparisons [29] was applied to determine whether the median percentage of each subset differed significantly (PB0.05) amongst the four age groups, whereas the Mann– Whitney U-test was used in the comparisons between the CD10 + strong and CD10 + weak groups. Differences in the distribution of the presenting clinical and biological features between the CD10 + strong and CD10 + weak groups were compared by the x 2 or Fischer exact test [29]. Life table estimates of DFS were derived by the method of Kaplan and Meyer and compared by the log-rank test [29,30]. Early death or failure to enter remission was considered an event at zero time. Considering the number of patients and the length of followup a difference in DFS of approximately 0.5 was expected to be detectable at an 80% confidence level both among the children and the adults. Univariate and multivariate analyses to determine significant prognostic factors were performed according to the Cox proportional-hazards model [30]. The clinical and biological features analyzed as covariates included: age ( B 10 vs. ] 10 years for children, and B35 vs. ] 35 years for adults), sex, race, presence of central nervous system (CNS) leukemia, presence of splenomegaly, hemoglobin level (B 10 vs. ] 10 g/dl), leukocyte count ( B 30.000/ml vs. ]30.000/ml), platelet count (B 20.000/ ml vs. ] 20.000/ml), FAB classification, presence of anomalous myeloid antigen, and CD10 intensity.

4. Results

4.1. Normal BM Table 1 shows the median percentage of CD19 + , CD19 + /CD10 + weak and CD19 + /CD10 + strong cells in the lymphocyte gate of BM biopsies from normal infants, children and adults. The B-subset represented more than half of the total lymphoid compartment in the BM from subjects younger than 15 years, and the percentage of CD19 + cells was significantly higher in the BM of children in the first 4 years of life. The CD19 + /CD10 + subset was the lymphoid subset most frequently found in BM from infants and children, its percentage decreasing with age. CD10 + weak cells predominated over CD10 + strong cells in all age groups, representing approximately 50% of the BM lymphoid cells in subjects younger than 4 years. The percentage of both CD10 + weak and CD10 + strong subsets was significantly lower in the BM from adults. The age-related decrease of the CD19 + /CD10 + weak and CD19 + / CD10 + strong subsets was balanced, resulting in a nonsignificant variation of the CD10 + weak/CD10 + strong ratio (Table 1). The CD10 and CD19 FI distribution in each CD10 subset did not change with age. Among age groups the median FI ranged from 14.1–16.3 afu in CD10 + weak cells and from 51.6–64.2 afu in CD10 + strong cells. The median CD19 FI ranged from 25–62.4 afu and was significantly higher in the CD19 + /CD10 + weak subset (median= 50 afu, p25th–p75th=29.8–64.6 afu) than in the CD19 + /CD10 + strong subset (median =31.2 afu, p25th–p75th= 17.8–45.5 afu)(P = 0.02).The highest FI detected in normal BM was of 110 afu for CD10 and of 100 afu for CD19. No significant difference in the CD10 FI was detected when CD2 instead of CD3 was used as T-marker (data not shown).

E.M. Rego et al. / Leukemia Research 23 (1999) 441–450

445

Table 2 Presenting clinical and biological features according to CD10 intensity in children with cALL Feature Age Gender Raceb Adenomegalyb Splenomegalyb

CNS leukemiab,c Hemoglobin

Leukocyte count

Platelets count

FAB classificationb,d Myeloid antigense

]10 years Male White Present Present ]5 cm B5 cm Present B7.5 g/dl ]7.5 and B10 g/dl ]10 g/dl B10.000/ml ]10.000/ml and B50.000/ml ]50.000/ml and B100.000/ml ]100.000/ml B20.000/ml ]20.000/ml and B100.000/ml ]100.000/ml L-1 Present

cALL CD10+strong (n =62)

cALL CD10+weak (n =16)

Pa

11.3% 59.7% 38/46 44/60 33/60 5/33 28/33 3/53 56.4% 30.6% 13% 50% 27.5% 8% 14.5% 22.6% 46.8% 30.6% 40/51 21%

31.2% 75% 12/13 12/16 15/16 9/15 6/15 1/13 62.5% 25% 12.5% 50% 31.3% 12.5% 6.2% 18.7% 56.3% 25% 10/13 6.3%

P =0.07 P =0.25 P= 0.36 P= 0.58 P= 0.04 P= 0.02 P= 0.6 P= 0.88

P= 0.80

P= 0.82

P=0.91 P= 0.11

Calculated by the x 2 or Fischer exact test. Results express the ratio: observed frequency/number of cases for which the data were available. c CNS, central nervous system. d All cases were classified either as L-1 or L-2. e Expression of CD33 or CD13 in an otherwise lymphoid phenotype. a

b

4.2. Clinical and biological features of patients with cALL CD10 + strong or CD10 + weak Tables 2 and 3 respectively show the presenting clinical and biological features of 78 children and 39 adults with cALL. A higher frequency of cases expressing the CD19 + /CD10 + strong phenotype was observed both in children and adults. Children in the cALL CD10 + weak group tended to be older than those in the cALL CD10 + strong group (median=7 years, p25th– p75th =3.8–10 years vs. median= 4 years, 3 – 6.5 years in the CD10 + strong group), but this difference was not significant. All children were older than 1 year, and eight of them were younger than 2 years, seven of whom were in the CD10 + strong group. In addition, a significantly higher frequency of splenomegaly was detected in the cALL CD10 + weak group. In about 60% of the cALL CD10 + weak cases, splenomegaly was massive ( ]5 cm below the costal margins) and this frequency was significantly higher than that found among children in the CD10 + strong group. No other clinical or biological difference was detected between children in each group. A significantly higher frequency of cases expressing the CD10 + weak phenotype was observed among adult women. Although the percentage of cases presenting with CNS leukemia was not significantly higher in the CD10 + weak group, the finding of this feature in three out of 11 patients in this group should be emphasized.

4.3. Treatment outcome The CRR for the 32 children analyzed was of 93.1% (S.D.= 9 25.8%), and did not differ significantly between the CD10 + weak (n=10, CRR= 98% 9 22.5%) and CD10 + strong groups (n= 22, CRR= 90.1% 9 29.4%). Fig. 1 shows the DFS for children with cALL, cALL CD10 + strong and cALL CD10 + weak. The estimated DFS was 68.9 months (95% confidence interval= 56.3–81.5 m) for all children, 67.9 months (95% CI= 50.9–84.9m) for the CD10 + strong group, and 57.8 m (95% CI= 40.6–75.1m) for the CD10 + weak group. For the 28 adults with cALL the CRR was 85.2% (9 36.2%) and no significant differences were observed between the CD10 + strong group (n= 18, CRR=77.89 42.8%) and the CD10 + weak group (n= 10, CRR =909 45%). Fig. 2 shows the DFS for these patients. The estimated mean DFS (95% IC) was 29.7 m (21–38.3 m) for all adults, 25.5 m (17.4–33.7 m) for the CD10 + strong group and 37.9 m (21.3–54.4 m) for the CD10 + weak. None of the features studied was an independent adverse prognostic feature for children or for adults, either by univariate or multivariate analysis.

4.4. Comparison of the CD10 and CD19 intensities in normal and leukemic BM Table 4 shows the CD10 and CD19 FI obtained for leukemic and normal B–cell progenitors. Leukemic

446

E.M. Rego et al. / Leukemia Research 23 (1999) 441–450

Table 3 Presenting clinical and biological features according to CD10 intensity in adults with cALL Feature

cALL CD10+strong (n =24)

cALL CD10+weak (n =15)

Pa P =0.46 P

Age Gender

B35 years Male

62.5% 75%

73.3% 26.7%

Raceb Adenomegalyb Splenomegaly

White Present Present ]5 cm B5 cm Present ]10 g/dl 530.000/ml \20.000/ml L-1 Present

18/20 13/22 58.3% 57.1% 42.9% 1/21 25% 62.5% 71.4% 7/19 4/19

9/12 7/12 46.7% 71.4% 28.6% 3/11 26.7% 57.1% 69.2% 5/11 2/9

CNS leukemiab,c Hemoglobin Leukocyte count Platelet count FAB classificationb,d Myeloid antigensb,e

= 0.001 P=0.26 P=0.62 P = 0.30 P =0.44 P= 0.11 P = 0.9 P= 0.57 P =0.89 P = 0.64 P=0.94

Calculated by the x 2 or Fischer exact test. Results express the ratio: observed frequency/number of cases for which the data were available. c CNS, central nervous system. d All cases were classified either as L-1 or L-2. e Expression of CD33 or CD13 in an otherwise lymphoid phenotype. a

b

CD10 + strong cells expressed significantly higher values of CD10 and CD19 FI than their normal counterparts. However, leukemic CD10 + weak cells expressed higher values of CD10 but not of CD19 intensity than normal BM progenitors. The FI of the CD10 and CD19 antigens for the normal and leukemic CD10 + weak groups, as well as for the normal and leukemic CD10 + strong groups are plotted in Fig. 3. The vertical and horizontal lines represent the highest values observed in normal BM for the CD10 and CD19 FI, respectively. The FI values above these (overexpression) were observed in 42.4% of the cALL cases. Of these, 29.6% overexpressed CD10, 7.7% overexpressed CD19 and 5.1% overexpressed both markers.

Fig. 1. Disease-free survival of children with cALL according to CD10 intensity. Legend: , cALL patients (n= 32); , cALL CD10 + strong patients (n = 22); , cALL CD10 + weak patients (n = 10).

In addition, the percentages of cells in normal marrow that exhibited CD10 FI higher than the cut off value of 110 afu and CD19 FI higher than 100 afu were of 0.1 and 0.5%, respectively. Moreover, only 0.01% of the normal cells expressed both antigens above these cut-off values.

5. Discussion The use of biopsies for the study of lymphoid subsets in BM is relevant for an accurate representation of tissue composition, excluding the effect of sample contamination with peripheral blood (PB) [28,31–33]. In a recent study we demonstrated that the B-lymphoid progenitors CD19 + CD10 + represent more than 50%

Fig. 2. Disease-free survival of adults with cALL according to CD10 intensity. Legend: , cALL patients (n =28); , cALL CD10 + strong patients (n =18); , cALL CD10 + weak patients (n =10).

E.M. Rego et al. / Leukemia Research 23 (1999) 441–450

447

Table 4 Comparison between the CD10 and CD19 fluorescence intensity in normal and leukemic B-cell progenitors Marker

Leukemic CD10+strong

Normal CD10+strong

Pb

Leukemic CD10+weak

Normal CD10+weak

CD10

90.9a (63.8–166.2)

62.3 (48.8–70)

PB0.001

19.9 (13.8–28.2)

15.3 (13.2–18.7)

P

50 (29.8–64.6)

=0.004 P =0.68

CD19 a b

48.6 (23.5–78.4)

31.2 (17.8–45.5)

P =0.001

50.6 (28.9–69)

Pb

Values are expressed as arbitrary fluorescence units. Calculated by the Mann–Whitney test.

of the lymphoid cells in BM biopsies from normal children younger than 4 years [28]. This percentage decreased with age, and in adult BM these cells represented 13% of all lymphoid cells. Moreover, in adults the percentages of T- and B-cells were approximately the same (median of 33.6% for CD19 + and of 34.8% for CD3 + cells). In the present study, we demonstrated that most of the normal B-cells in BM express the CD10 antigen weakly, but a small percentage of CD10 + strong cells was detected in all age ranges. Although the predominance of the CD10 + weak subpopulation over the CD10 + strong one was demonstrated in previous investigations [1,2,34], this is the first study to assess age-related differences. The CD10 + weak/ CD10 + strong ratio ranged from 12.6 to 25.9, without a significant difference amongst the age groups, demonstrating that both CD10 + subpopulations decreased with age in a balanced way. We did not detect any significant difference in CD10 or CD19 fluorescence intensity with age, in contrast with the reported pattern of other markers [35]. Moreover, CD19 expression was stronger in CD10 + weak cells in agreement with Farahat et al. [6], who demonstrated that normal lymphoid TdT − precursors presented lower CD10 and higher CD19 expression than TdT + precursors. Different techniques have been previously used to assess the intensity of expression of an antigen [36–40]. In this study, we decided to express the intensity of antigenic expression by the ratio between the mean fluorescence channel of the sample and that of normal T-cells. We considered this a more reproducible approach than the use of background fluorescence as a control. In addition, similar results were obtained when T-cells were identified by the expression of CD2 or CD3. Although this is a simple and less costly method for performing this quantification, it requires careful standardization of instrument and reagents. Another concern is the criteria used for the definition of the CD10 + weak and CD10 + strong groups. Although the choice of the cut-off value of 35 was arbitrary, it was based on the observation that in the majority of BM samples from normal children the fluorescence histogram presented a dip around this value separating two subpopulations according to CD10 intensity.

Based on previous reports demonstrating differences in maturational stage [1,2,4,34,41], in proliferative activity [5,42], and in the pattern of cytogenetic abnormalities [7] between CD10 + weak and CD10 + strong B-precursors, we proposed to study the clinical profile and the treatment outcome of patients, whose blasts expressed the CD10 strongly or weakly. Most cALL cases expressed the CD19 + CD10 + strong phenotype both in children and adults. Although different methods were employed this result is in agreement with that obtained by Lavabre-Bertrand et al. [7]. We observed that children in the CD10 + weak group tended to be older and presented a significantly higher frequency of splenomegaly, which in the majority of cases was massive. In the present study the CD10 FI was not related to sex or race in contrast with the results reported by Glencross et al. [8]. However, the epidemiological features of ALL of African and Brazilian children are distinct [21,22]. A significantly higher frequency of females was observed in the adult CD10 + weak group. In addition, the frequency of CNS leukemia was higher in the CD10 + weak group (statistically nonsignificant), being above that described in the literature for adults with ALL [17,43]. However, considering the small size of the sample, both findings must be further confirmed. Unfortunately, in the present study the treatment outcome analyses were affected by the small sample size. However, the observed CCR and DFS amongst children were similar to those reported for larger number of patients using equivalent protocols [25]. Amongst adults, the CRR was similar, but the DFS was lower than the values reported in the literature [26,27,44]. Since cytogenetic analyses of the cases were not performed, it is probable that a considerable number of Ph1 positive cases (known to have a poorer prognosis) were included. Moreover, this cytogenetic abnormality is particularly common among adults with cALL [17,45]. The estimated CCR and DFS for the CD10 + weak and CD10 + strong groups did not differ significantly for children or for adults, and CD10 FI was not of prognostic significance. These results contrast with those obtained by Look et al. [9], possibly owing to the important methodological differences between the two studies. On the other hand, our results agree

448

E.M. Rego et al. / Leukemia Research 23 (1999) 441–450

Fig. 3. CD10 and CD19 intensities in normal and leukemic B-cell precursors. Legend: , normal B-precursors CD10 + strong; , normal B-precursors CD10 + weak; , leukemic B-precursors CD10 + strong; , leukemic B-precursors CD10 + weak. The vertical line indicates the highest intensity of CD10 observed in normal BM, and the horizontal line the highest intensity of CD19 observed in normal BM. *, Percentage of the cALL cases.

with more recent studies in which the expression of CD10 was found to lack prognostic significance in cALL [10,13]. The comparison of the CD10 and CD19 fluorescence intensity between normal and leukemic lymphoid precursors revealed that CD10 expression was stronger in leukemic cells than in normal precursors, both in CD10 + strong and CD10 + weak cells. The CD19 expression was also stronger only in CD10 + strong leukemic B-progenitors. These results are in agreement with those obtained by Farahat et al. [6] using quantimetry, showing that TdT + leukemic cells expressed a higher number of CD10 and CD19 molecules than their normal counterparts. Similarly, in about 40% of cALL cases we demonstrated a lack of the expected inverse relationship of CD10 and CD19 expression. In this study, we demonstrated the relevance of the CD10 and CD19 FI analysis to the detection of minimal residual leukemia in 40% of cALL cases, where CD10 was expressed in intensities higher than 110 afu and/or CD19 at intensities higher than 100 afu. For these cases, the leukemic cells were detectable at levels

of 0.2% based on the CD10 analysis, and of 0.6% based on the CD19. More relevant is the overexpression of both antigens, in which cases leukemic blasts were detectable at levels of 0.02%. The fact that most lymphoid cells in normal children’s BM have the same immunophenotype as the leukemic cells in cALL shows that special caution should be exercised when samples from children are analyzed. In conclusion, in the present study we demonstrated that the heterogeneity of CD10 and CD19 intensity is of no clinical relevance in cALL, but its analysis may be helpful for the diagnosis and for the detection of minimal residual disease in about 40% of cALL cases.

Acknowledgements The authors gratefully acknowledge the assistance of Professor J Carneiro and the cardiac surgery staff in obtaining the sternal biopsies, the assistance of Gerson Muccilo in the statistical analysis and of Elettra Greene

E.M. Rego et al. / Leukemia Research 23 (1999) 441–450

in preparing this manuscript. This study was supported by Grant No. 520.786/96-3 from the Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gicoCNPq. E.M. Rego contributed to the concept, design, analysis of data, drafting of the article, critical revision and final approval and data collection. L.G. Tone provided study materials. A.B. Garcia provided technical support and collection of data. R.P. Falcao contributed to the concept and design, critical review and final approval, provision of study materials, obtaining of funding and administrative support.

References [1] Ryan D, Kossover S, Mitchell S, Frantz C, Hennessy L, Cohen H. Subpopulations of common acute lymphoblastic leukemia antigen-positive lymphoid cells in normal bone marrow identified by hematopoietic differentiation antigens. Blood 1986;68:417. [2] Loken MR, Shah VO, Dattilio KL, Civin CI. Flow cytometric analysis of human bone marrow. II. Normal B-lymphocyte development. Blood 1987;70:1316. [3] Janossy G, Bollum FJ, Bradstock KF, Ashley J. Cellular phenotypes of normal and leukemic hemopoietic cells determined by analysis with selected antibody combinations. Blood 1980;56: 430. [4] Ryan DH, Chapple CW, Kossover SA, Sandberg AA, Cohen HJ. Phenotypic similarities and differences between CALLApositive acute lymphoblastic leukemia cells and normal marrow CALLA-positive B-cell precursors. Blood 1987;70:814. [5] Campana D, Janossy G. Proliferation of normal and malignant human immature lymphoid cells. Blood 1988;71:1201. [6] Farahat N, Lens D, Zomas A, Morilla R, Matutes E, Catovsky D. Quantitative flow cytometry can distinguish between normal and leukaemic B-cell precursors. Br J Haematol 1995;91:640. [7] Lavabre-Bertrand T, Janossy G, Ivory K, Peters R, SeckerWalker L, Porwit-MacDonald A. Leukemia-associated changes identified by quantitative flow cytometry: I. CD10 expression. Cytometry 1994;18:209. [8] Glencross DK, Adam F, Poole J, Cohn R, Becker P, Fleming AF, Mendelow BV. CD10 antigen density in childhood common acute lymphoblastic leukaemia: comparisons of race and sex. Leuk Res 1992;16:1197. [9] Look AT, Melvin SL, Brown K, Dockter ME, Roberson PK, Murphy SB. Quantitative variation of the common acute lymphoblastic leukemia antigen (gp 100) on leukemic marrow blasts. J Clin Investig 1984;73:1617. [10] Pui C-H, Rivera GK, Hancock ML, Raimondi SC, Sandlund JT, Mahmoud HH, Ribeiro RC, Furman WL, Hurwitz CA, Crist WM, Behm FG. Clinical significance of CD10 expression in childhood acute lymphoblastic leukemia. Leukemia 1993;7:35. [11] Greaves MF, Janossy G, Peto J, Kay H. Immunologically defined subclasses of acute lymphoblastic leukemia in children: their relationship to presentation features and prognosis. Br J Haematol 1981;48:179. [12] Pui C-H, Williams DL, Raimondi SC, Melvin SL, Behm FG, Look AT, Dahl GV, Rivera GK, Kalwinsky DK, Mirro J, Dodge RK, Murphy SB. Unfavorable presenting clinical and laboratory features are associated with CALLA-negative non-T, non-B-lymphoblastic leukemia in children. Leuk Res 1986;10:1287. [13] Pui C-H, Behm FG, Crist WM. Clinical and biological relevance of immunologic marker studies in childhood acute lymphoblastic leukemia. Blood 1993;82:343.

449

[14] Pullen DJ, Boyett JM, Crist WM, Falletta JM, Roper M, Dowell B, van Eys J, Jackson JF, Humphrey GB, Metzgar RS, Cooper MD. Pediatric Oncology Group utilization of immunologic markers in the designation of acute lymphocytic leukemia subgroups: influence on treatment response. Ann N Y Acad Sci 1984;428:26. [15] Crist W, Pullen J, Boyett J, Falletta J, Van Eys J, Borowitz M, Jackson J, Dowell B, Frankel L, Quddus F, Ragab A, Vietti T. Clinical and biological features predict a poor prognosis in acute lymphoid leukemia in infants. A Pediatric Oncology Group study. Blood 1986;67:135. [16] Vannier JP, Bene MC, Faure GC, Bastard C, Garand R, Bernard A. Investigation of the CD10 (cALLa) negative acute lymphoblastic leukemia: further description of a group with poor prognosis. Br J Haematol 1989;72:156. [17] Copelan EA, McGuire EA. The biology and treatment of acute lymphoblastic leukemia in adults. Blood 1995;85:1151. [18] Bene MC, Faure GC. CD10 in acute leukemias. GEIL (Groupe ´ tude Immunologique des Leucemies). Haematologica dE 1997;82:205. [19] FAB Cooperative Group, Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DAG, Gralnick HR, Sultan C. Proposals for the classification of the acute leukemias. Br J Haematol 1976;33:451. [20] FAB Cooperative Group, Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DAG, Gralnick HR, Sultan C. The morphological classification of the acute lymphoblastic leukemia: concordance among observers and clinical correlation. Br J Haematol 1981;47:553. [21] Greaves MF, Colman SM, Beard MEJ, Bradstock K, Cabrera ME, Chen PM, Jacobs P, Lam-Po-Tang PRL, MacDougall LG, Williams CKO, Alexander FE. Geographical distribution of acute lymphoblastic leukemia subtypes: second report of the collaborative group study. Leukemia 1993;7:27. [22] Rego EM, Garcia AB, Viana SR, Falca˜o RP. Characterization of acute lymphoblastic leukemia subtypes in Brazilian patients. Leuk Res 1996;20:349. [23] Buccheri V, Matutes E, Dyer MG, Catovsky D. Lineage commitment in biphenotypic acute leukemia. Leukemia 1993;7:919. [24] Matutes E, Catovsky D. The value of scoring systems for the diagnosis of biphenotypic leukemia and mature B-cell disorders. Leuk Lymphoma 1994;13(Suppl. 1):11. [25] Brandalise S, Odone V, Pereira W, Andrea M, Zanichelli M, Aranega V. Treatment results of three consecutive Brazilian cooperative childhood ALL protocols: GBTLI-80, GBTLI-82 and -85 ALL Brazilian Group. Leukemia 1993;7(Suppl. 2):142. [26] Clarkson B, Ellis S, Little C, Gee T, Arlin Z, Mertelsmann R, Andreeff M, Kempin S, Koziner B, Chaganti R, Jhanwar S, McKenzie S, Cirrincione C, Gaynor J. Acute lymphoblastic leukemia in adults. Semin Oncol 1985;12:160. [27] Clarkson B, Gaynor J, Little C, Berman E, Kempin S, Andreeff M, Gulati S, Cunningham I, Gee T. Importance of long-term follow up in evaluating treatment regimens for adults with acute lymphoblastic leukemia. Hamatol Bluttransfus 1990;33:397. [28] Rego EM, Garcia AB, Viana SR, Falca˜o RP. Age-related changes in lymphocyte subsets in normal bone marrow biopsies. Cytometry 1998;34:22. [29] Armitage P, Berry G. Statistical methods in medical research, 3rd edition. Oxford: Blackwell, 1994. [30] Friedman LM, Furberg CD, DeMets DL. Fundamentals of clinical trials, 3rd edition. St. Louis: Mosby, 1996. [31] Clark P, Normansell DE, Innes DJ, Hess CE. Lymphocyte subsets in normal bone marrow. Blood 1986;67:1600. [32] Fauci AS. Human bone marrow lymphocytes. I Distribution of lymphocyte subpopulations in the bone marrow of normal individuals. J Clin Investig 1975;56:98.

450

E.M. Rego et al. / Leukemia Research 23 (1999) 441–450

[33] Gale RP, Opelz G, Kiuchi M, Golde DW. Thymus-dependent lymphocytes in human bone marrow. J Clin Investig 1975;56:1491. [34] Dworzak MN, Fritsch G, Fleischer C, Printz D, Froschl G, Buchinger P, Mann G, Gadner H. Multiparameter phenotype mapping of normal and post-chemotherapy B-lymphopoiesis in pediatric bone marrow. Leukemia 1997;11:1266. [35] Bikoue A, George F, Poncelet P, Mutin M, Janossy G, Sampol J. Quantitative analysis of leukocyte membrane antigen expression: normal adult values. Cytometry 1996;26:137. [36] Goldmacher VS, Lambert JM, Young AY, Anderson J, Tinnel NL, Kornacki M, Ritz J, Blatter WA. Expression of the common acute lymphoblastic leukemia antigen (cALLa) on the surface of individual cells of human lymphoblastoid lines. J Immunol 1986;136:320. [37] Poncelet P, Carayon P. Cytometric quantification of cell-surface antigens by indirect immunofluorescence using monoclonal antibodies. J Immunol Methods 1985;85:65. [38] Vogt Jr RF, Cross GD, Henderson LO, Phillips DL. Model system evaluating fluorescein-labeled microbeads as internal standards to calibrate fluorescence intensity on flow cytometers. Cytometry 1989;10:294.

.

[39] Caldwell CW, Maggi J, Henry LB, Taylor HM. Fluorescence Intensity as a quality control parameter in clinical flow cytometry. Am J Clin Pathol 1987;88:447. [40] Caldwell CW, Patterson WP. Fluorescence intensity of immunostained cells as a diagnostic aid in lymphoid leukemias. Diagn Clin Immunol 1988;5:371. [41] Caldwell CW, Poje E, Helikson MA. B-cell precursors in normal pediatric bone marrow. Am J Clin Pathol 1991;95:815. [42] Hollander Z, Shah VO, Civin CI, Loken MR. Assessment of proliferation during maturation of the B-lymphoid lineage in normal human bone marrow. Blood 1988;71:528. [43] Hoelzer D. Diagnosis and treatment of adult acute lymphocytic leukemia. In: Wiernik PH, Canellos GP, Kyle RA, Schiffer CA, editors. Neoplastic diseases of the blood. New York: Churchill Livingstone, 1991:253. [44] Hoelzer D. Acute lymphoblastic leukemia in adults. In: Henderson ES, Lister TA, Greaves MF, editors. Leukemia. Philadelphia: Saunders, 1996:446. [45] LoCoco F, Foa R. Diagnostic and prognostic advances in the immunophenotypic and genetic characterization of acute leukaemia. Eur J Hematol 1995;55:1.