(Acta Paediatr J m 1991: 33: 507
- 521)
Immunological Classification of Childhood Acute Lymphoblastic Leukemia Shinpei Nakazawa, M.D.’, Midori Saito, M.D.2, Toshiko Okazaki, M.D.3, Keiko Takane, M.D.3, Kanji Sugita, M.D.*, Taijiro Mori, M.D.*, Kazuyoshi Nishino, M.D.2, Toshio Suzuki, M.D.*, Akitoshi Kinoshita, M.D.*, Takashi Abe, M.D.2, Yoshihiro Kurosawa, M.D.2 and Takeshi Inukai, M.D.’ ‘Department of Pediatrics, Yamanashi Medical College, Yamanashi; 2Department of Pediatrics, School of Medicine, Keio University, Tokyo; 31mmunology Research Laboratory, Social Insurance Saitama Central Hospital, Saitama
Seven hundred and forty-four newly diagnosed patients with acute leukemias between 1978 and 1990 were classified on the basis of immunological phenotypes. The majority of the patients were enrolled in the Tokyo Children’s Cancer Study Group (TCCSG) studies. The incidence of subclassification of acute leukemias in this study was as follows: 522 patients with ALL (70%), 139 patients with ANLL (18%), 29 patients with biphenotypic leukemia, 8 patients with Phl-positive acute leukemia (Ph’-AL), and 45 patients with infant leukemia. ALLs were classified into common ALL (CALL, 77%), T-ALL (15%), B-ALL (4%), and unclassified ALL (3%). The incidence of ALL subtypes in this study reflected those of TCCSG. Biphenotypic leukemias were categorized into 4 groups as follows; 1) CALL with positive myelomonocytic antigen@) (N = ll), 2) unclassified ALL with positive myelomonocytic antigen(s) (N = 5), 3) ANLL with positive B-lymphoid antigen(s) (N = 4), and 4) acute leukemia with positive T-lymphoid and myeloid antigen(s). Infant leukemias were classified into ALL type (N=27) and ANLL type (N=18). In this present study, clinical features and immunological phenotypes of the acute leukemias with a poor prognosis, i.e. biphenotypic leukemia, Phl-AL, and infant leukemia are analyzed and discussed. Key Words
Childhood ALL, Biphenotypic leukemia, Ph’ positive acute leukemia, Immunological marker, Cytoplasmic CD13 antigen
Introduction
Received July 5. 1991 Correspondence address: Shinpei Nakazawa, M.D., Department of Pediatrics, Yamanashi Medical College. 1 110, Shimokato, TamahMho, Nakakomagun, Y m a nashi 409-38, Japan
Since cell surface markers of human T- and B-lymphocytes were recognized in the early 1970’s, the immunological classification of childhood acute lymghoblastic leukemia (ALL) is not Only to basic investigations, but also to the stratification of ALL
508 (88) Nuku:u~*uel al. treatment and the evaluation of the treatment results. During the 1970's- surface immunoglobulins (slg), rosette formation with sheep red blood cells and T-antigen(s) recognized by heteroantiserum were the major markers for classification. Additionally, early studies by Greaves et al demonstrated the existence of common acute lymphoblastic leukemia antigen (CALLA) in the majority of childhood ALLs [I]. By these markers, ALLs have been classified into 3 categories, i.e. common ALL, T-ALL, and B-ALL. This classification is still used. Since the monoclonal antibody technology was established in the 1980's, monoclonal antibodies which recognize the antigens on hematopoietic cells have become available for the diagnosis and classification of leukemias. At present, the monoclonal antibodies are categorized in 78 clusters (CD) described in the reports of the First to Fourth International Workshops on Leukocyte Differentiation Antigens 12-51. In this report, the historical data of immunological classification of ALL and its clinical significance during the last decade experienced in our laboratory are described. Additionally, the results of cell surface marker analysis of miscellaneous types of acute leukemias, i.e. biphenotypic leukemia, infant leukemia, and Phi positive acute leukemia are reported.
Materials and Methods Patients Seven hundred and forty-four newly diagnosed patients with acute leukemias were studied between 1978 and 1990. Among these patients, six hundred and ninety-nine patients (949) were between 1 and 15 years old and forty-five patients (6%) were between 1 day and 12 months old. The majority of the patients were enrolled into the treatment protocols of the Tokyo Children's Cancer Study Group (TCCSG) [6].
Cell separation and conventional hematological studies The heparinized bone marrow blood or peripheral blood were transferred to our laboratory for immunological studies from our medical department, affiliated hospitals of Keio University and the medical institutes contributing to the TCCSG study. The percentages of blasts in the samples were shown to exceed 60% in all 744 patients. Bone marrow or peripheral blood smears were stained by the May-Giemsa method. Conventional cytochemical stainings, such as myeloperoxidase, specific and non-specific esterase, Sudan-black, and periodic acid Schiff reaction were also performed. Mononuclear cells were separated by Hypaque-Ficoll density gradient centrifugation. After the cell separation, mononuclear cells were washed and suspended in phosphate buffer saline solution for immunological studies. In some cases, the separated cells were suspended in RPMI 1640 media supplemented with 20% fetal bovine serum and cultured without any mitogens or cytokines for 3 days in 5% CO, atmosphere (short-term culture). Rosette formation Rosette formation with sheep red blood cells (E-rosette) and erythrocyte-complementcomplex (EAC-rosette) were carried out by the method previously described [7]. Surface and cytoplasmic immunoglobulins Cell surface immunoglobulins (a, y. p, K and A chains) were determined by direct immunofluorescent staining using rabbit antia, y. p, K and h chains sera (Hoechst, Behringwerke, Germany). Cytoplasmic p chains were detected by the double peroxidase-antiperoxidase (PAP) method described elsewhere [81. Cell surface antigens From 1978 to 1980, indirect immunofluorescent staining was carried out for detecting cell surface antigens. The antibodies used in this period were well characterized rabbit anti-T cell, anti-CALLA and HLA-DR antisera
Actu
Puediutr Jpn
Inirnunological classific,ationof' A L L (89) 509
produced by the method previously described [9-111. From 1981, monoclonal antibodies were used by the micro-rosette method and flow cytometry. The micro-rosette method was developed by our laboratory [ 121 through the modification of previous methods [13, 141 for analyzing small numbers of cells. Briefly, the separated mononuclear cells were allowed to adhere on a Terasaki microtest plate (Falcon, #3034, USA) treated with po1y-Llysine. After the neutralization of positive charge of poly-L-lysine by fetal bovine serum, an adequate amount of monoclonal antibody was gently added, washed and ox red blood cells conjugated with goat-anti-mouse IgG antibody were added. After the reaction between the adhered mononuclear cells and the ox red blood cells, the microtest plate was placed upsidedown to remove non-reactive red cells from the bottom of the plate and the rosette forming cells were counted by light microscopy. The sensitivity of this method was compatible with those of indirect immunofluorescent staining. The clusters defined by monoclonal antibodies used in this study were CD6 (T6), CD2 (Tll), CD3 (T3), CD4 (T4), CD5 (TI), CD7 (Tp40), CD8 (T8) for T-lymphoid antigens, CDlO (J5, KOR-N34 [IS]), CD19 (B4), CD20 (BI), CD22 (Leul4) for B lymphoid antigens, C D l l (Leu15), CD13 (MY7, MCS2), CD14 (MY4), CD33 (MY9), CD36 (OKM5) for myelomonocytic antigens and HLA-DR antigen (KOR-la17 [15], 12, anti-HLA-DR). Heteroantisera (anti-T, anti-la, and anti-CALLA) produced in our laboratory were also utilized. In the late 1980's, CD41 (TP80), CD42 (AN51), CD56 (NKHI) and KOR-P77 [I61 (anti-platelet antigen) were added to the antibody panel and PFI, GTCSI were used for T-cell type ALL in recent studies. The reaction of the above antigens was defined as positive when antigen(s) were expressed on 30% or more cells determined by microrosette method or flow cytometry (Epics-CS, Coulter, USA). When blasts expressed both lymphoid and myelomonocytic antigens, double staining was performed.
Vol. 33 No. 4 August 1991
Cytoplasmic antigen From 1988, cytoplasmic antigens were determined by the double PAP method [8]. The antibodies used in this study were MY7 (CD13), T3 (CD3), P F l , GTCSI. The blasts from 277 patients with common ALL, 17 patients with unclassified ALL and some patients with biphenotypic leukemia were analyzed by using CD13. Cytoplasmic T-cell receptor antigens defined by T3, PFl, and GTCSl were examined in blasts from T-ALL and biphenotypic leukemias with T-cell marker( s). Diagnostic criteria The definition of ALL was based on FAB classification, conventional cytochemistry and the expression of lymphoid antigen(s) on the cell surface. Common ALL was diagnosed by the expression of CDIO, CD19, and/or CD22 and HLA-DR. Pre-B ALL was defined by the existence of cytoplasmic p chain in addition to the CDlO and CD19. The diagnostic markers of T-ALL and B-ALL were the expression of T-lymphoid antigens and E-rosette formation, and B-lymphoid antigens and surface immunoglobulins, respectively. Unclassified ALL was tentatively defined in terms of the expression of B-lymphoid antigens (CD19) and the loss of CDIO, T-lymphoid antigens and myelomonocytic antigens. In this study a diagnosis of biphenotypic leukemia was tentatively made when the expression of myeloid antigen(s) on the cell surface was proved in addition to lymphoid antigens. In this study patients with acute leukemia under 1 year were categorized as infant leukemia because of the different clinical features and poor prognosis. Ph' positive acute leukemia (Phi-AL) was diagnosed by the demonstration of chromosomal translocation t(9;22)(q34;qll) and bcr-abl chimeric mRNA. Cytogenetic study Cytogenetic analysis was performed in some patients from the late 1980's. G-banding and Q-banding methods were performed for chromosomal analysis in this study.
5 10 (90) N i r k u z ~ ~et~al. ~u
DNA and RNA analysis DNA and mRNA were extracted from the leukemic cells of patients with biphenotypic leukemias and infant leukemias. The rearrangement of T-cell receptor gene (Cp) and immunoglobulin heavy chain gene (JH) were studied by the standard Southern blotting method. Bcr-abl kimeric mRNA was defined by reverse transcriptase polimerase chain reaction (PCR) by using the following primers, major bcr exon 2 (b2) and 3 (b3), minor bcr exon 1 ( b l ) and abl exon 2 (a2), of which the characteristics have been described elsewhere [ 171.
Table I . The incidence of acute leukemia Patient number
522
ALL Common ALL T-ALL B-ALL Unclassified ALL
ANLL Biphenotypic leukemia Ph'-positive acute leukemia NK-leukemia Infant leukemia
404 (77.4%) 81 (15.5%) 21 (4.0%) 16 (3.1%) 139 29 8 1 45
A l 1.: acute Iymphoblautic leukemia. ANLL.: acute noii-l!mphocytic leubemia. N K : natural killer.
Results Incidence of ALL subtypes The overall incidence of subtypes of acute leukemias analyzed in this study is shown in Table 1. In 522 patients with ALL, 404 (77.4%). 81 (15.5%), 21 (4.0%) and 16 (3.1%) patients were diagnosed as common ALL, TALL. B-ALL and unclassified ALL, respectively. The incidence and prognosis of pre-B ALL could not be determined because of the small number in analyzed samples. Our preliminary study demonstrated that the 12 out of 44 common ALL (27.3%) which had cytoplasmic /.I chain, were not different from that of common ALL in an event free survival (EFS). Ph'-AL could not be differentiated from common ALL before the 1986's. In all seventeen patients with unclassified ALL, only CD19 antigen was expressed and other lymphoid antigens could not be detected. The ALL'ANLL ratio (4:I ) does not reflect the true incidence in the TCCSG study, since the blood samples from patients with ANLL were not transferred before the middle of the 1980's. There was one patient with blasts expressing natural killer activity and CD56 (NKH-I). The marker profiles and clinical features have been described elsewhere [ 181. Common ALL Common ALL is the most common subtype among childhood acute leukemias. The majority of common ALL blasts expressed
B-lineage antigens (CD10, CD19, CD22 and HLA-DR). In this study the expression of cell surface CD13 antigen was examined in 277 patients with common ALL. In 10 out of 277 cases, blasts expressed CD13. These 10 cases were categorized as biphenotypic leukemia. Cytoplasmic CD13 (cCD13) was examined in 59 newly diagnosed common ALL cases in which the expression of surface CD13 was negative. In 11 out of 59 patients (18.6%), cCD13 antigen was demonstrated. After short term culture without any mitogens or cytokines, surface CD13 appeared in all 11 cases. With the extensive studies including karyotype analysis and mRNA by PCR, five out of these 1 I patients were categorized as Ph'-AL based on the existence of t(9;22) (q34; q l l ) translocation and bcr-abl kimeric mRNA. T-ALL In an earlier study, heteroantiserum and Erosette formation were used for the diagnosis of T-ALL, since the monoclonal antibodies were not available. Thus these cases were reexamined by monoclonal antibodies using recovered leukemic cells which were stored in liquid nitrogen and the diagnosis was reconfirmed. Fig. I shows the incidence of cell surface and cytoplasmic antigens in 42 patients with T-ALL. In all cases, CD7 and cytoplasmic CD3 (cCD3) were expressed. Table 2 shows the proposal of the subclassification of these 42 patients based on the
A d a Paediatr Jprz
Inimunological classi$carion of' A L L (91) 51 I
sCD sCD sCD sCD sCD sCD
0
20
40
60
80
100 ( 0 4
Positive Cases Fig. I : Incidence of surface and cytoplasmic T-antigens in T-ALL. CDI, 3. 4, 8, 5 , 2. 7 were defined on the cell surface by micro-rosette method. T-ALL: Forty-two cases were well characterired for cell surface and cytoplasmic antigens. SFI : recogniie TcRP chain.
expression of T-cell receptor (TcR) antigens. The most common pattern was cbTCS1-, cPFI+, sCD3- subtype. The clinical significance of this classification is now under investigation. In this study, further classifications based on other T-cell differentiation antigens were tried (data not shown). The pattern of antigen expression was heterogeneous and it was impossible to divide T-ALLs into subgroups as previously described [19, 201. B-ALL The blasts in B-ALL expressed cell surface immunoglobulin, CDIO, CD19, CD20, and CD22. The most common immunoglobulin subtype was IgM @). The morphology of these were FAB-L3 (18 cases) and L1 (3 cases). There were some cases indistinguishable from Burkitt's lymphoma at stage IV. B-ALL had a poor prognosis even if intensive chemotherapy was employed. The blasts with LI morphology (3 cases) displayed a small amount of Y, chain on the cell surface [21]. Biphenotypic leukemias Although the definition, nomenclature and diagnostic criteria of biphenotypic leukemia
Vol. 33 No. 4 August 1991
Table 2. Subclassification of T-ALL based on T-cell receptor antigens d T C S I i cBFl i sCD3
-i-I+I-i-I+--I+ or - 1 -
Incidence (%)
8 (19%) 1(2%) 23 (55%) 10 (24%)
T-AI.l.: Forty-two cases were well characteriied for cell surface and cytoplasmic antigens. c6TCS I : cytoplasmic TcK 6 chain recognized by GTCSI. cPFI: cytoplasmic TcR chain recogni7ed by PFI. sCD3: cell surface CD3 antigen recogniied by T3.
have not been established, the diagnosis of biphenotypic leukemia was tentatively made in this study when one of the following criteria was fulfilled: (1) the coexpression of both lymphoid and myelomocytic antigen(s) on the same leukemic cells (biphenotypic), (2) coexistence of the cells expressing different cell lineage characteristics (lymphoid and myelomonocytic) and the cells were proved to have the same clonal origin, and (3) the cell lineage changed during the course of the disease (lineage switch). Twenty-nine out of 744 patients were diagnosed as biphenotypic leukemia. These patients were divided into 4
5 I2 (92) Naliu:urru Pt at.
myeloid antigen(s) were confirmed by double staining on flow cytometry. The antigenic markers of these cases are summarized in Tables 3 to 7. In 11 patients with common ALL (group I), the leukemic cells expressed CD13 (lo/ 11). In one case (patient #ll), CD36 was expressed in addition to CDlO, CD19 and HLA-DR. In five patients with unclassified ALL (group 2), heterogeneity of
groups as follows: ( 1 ) common ALL with positive myelomonocytic antigen(s) expression, (2) unclassified ALL with positive myelomonocytic antigen(s) expression, (3) ANLL (Ml) with positive B- lymphoid antigen(s) expression and (4) acute leukemia with positive T- and myeloid antigen(s) expression (T/M hybrid leukemia). In the majority of these cases, biphenotypic expression of lymphoid and
Table 3. Group I , common ALL with positive myelomonocytic antigen(s) expression Patient no.
Age (Yr)
WBC
1
15600
'1.
50800 3800 I600 7400 8800 3900 5900 32600 I300 139900
I. 2. 3. 4. 5. 6.
4 4
7. 8. 9. 10. II .
8 9 II 12 14
S
5
(:uU
HLA-DR (12)
CDlO (JS)
CD19
(B4)
CD20 (BI)
86 93 88 96 43 92 85 97 86 100 87
87 98 64 96 98 92 83 80 83 96 80
nt
nt
98 83 9s nt 96 84 nt 82 nt 91
17 43 53 nt
- 521)
Immunological Classification of Childhood Acute Lymphoblastic Leukemia Shinpei Nakazawa, M.D.’, Midori Saito, M.D.2, Toshiko Okazaki, M.D.3, Keiko Takane, M.D.3, Kanji Sugita, M.D.*, Taijiro Mori, M.D.*, Kazuyoshi Nishino, M.D.2, Toshio Suzuki, M.D.*, Akitoshi Kinoshita, M.D.*, Takashi Abe, M.D.2, Yoshihiro Kurosawa, M.D.2 and Takeshi Inukai, M.D.’ ‘Department of Pediatrics, Yamanashi Medical College, Yamanashi; 2Department of Pediatrics, School of Medicine, Keio University, Tokyo; 31mmunology Research Laboratory, Social Insurance Saitama Central Hospital, Saitama
Seven hundred and forty-four newly diagnosed patients with acute leukemias between 1978 and 1990 were classified on the basis of immunological phenotypes. The majority of the patients were enrolled in the Tokyo Children’s Cancer Study Group (TCCSG) studies. The incidence of subclassification of acute leukemias in this study was as follows: 522 patients with ALL (70%), 139 patients with ANLL (18%), 29 patients with biphenotypic leukemia, 8 patients with Phl-positive acute leukemia (Ph’-AL), and 45 patients with infant leukemia. ALLs were classified into common ALL (CALL, 77%), T-ALL (15%), B-ALL (4%), and unclassified ALL (3%). The incidence of ALL subtypes in this study reflected those of TCCSG. Biphenotypic leukemias were categorized into 4 groups as follows; 1) CALL with positive myelomonocytic antigen@) (N = ll), 2) unclassified ALL with positive myelomonocytic antigen(s) (N = 5), 3) ANLL with positive B-lymphoid antigen(s) (N = 4), and 4) acute leukemia with positive T-lymphoid and myeloid antigen(s). Infant leukemias were classified into ALL type (N=27) and ANLL type (N=18). In this present study, clinical features and immunological phenotypes of the acute leukemias with a poor prognosis, i.e. biphenotypic leukemia, Phl-AL, and infant leukemia are analyzed and discussed. Key Words
Childhood ALL, Biphenotypic leukemia, Ph’ positive acute leukemia, Immunological marker, Cytoplasmic CD13 antigen
Introduction
Received July 5. 1991 Correspondence address: Shinpei Nakazawa, M.D., Department of Pediatrics, Yamanashi Medical College. 1 110, Shimokato, TamahMho, Nakakomagun, Y m a nashi 409-38, Japan
Since cell surface markers of human T- and B-lymphocytes were recognized in the early 1970’s, the immunological classification of childhood acute lymghoblastic leukemia (ALL) is not Only to basic investigations, but also to the stratification of ALL
508 (88) Nuku:u~*uel al. treatment and the evaluation of the treatment results. During the 1970's- surface immunoglobulins (slg), rosette formation with sheep red blood cells and T-antigen(s) recognized by heteroantiserum were the major markers for classification. Additionally, early studies by Greaves et al demonstrated the existence of common acute lymphoblastic leukemia antigen (CALLA) in the majority of childhood ALLs [I]. By these markers, ALLs have been classified into 3 categories, i.e. common ALL, T-ALL, and B-ALL. This classification is still used. Since the monoclonal antibody technology was established in the 1980's, monoclonal antibodies which recognize the antigens on hematopoietic cells have become available for the diagnosis and classification of leukemias. At present, the monoclonal antibodies are categorized in 78 clusters (CD) described in the reports of the First to Fourth International Workshops on Leukocyte Differentiation Antigens 12-51. In this report, the historical data of immunological classification of ALL and its clinical significance during the last decade experienced in our laboratory are described. Additionally, the results of cell surface marker analysis of miscellaneous types of acute leukemias, i.e. biphenotypic leukemia, infant leukemia, and Phi positive acute leukemia are reported.
Materials and Methods Patients Seven hundred and forty-four newly diagnosed patients with acute leukemias were studied between 1978 and 1990. Among these patients, six hundred and ninety-nine patients (949) were between 1 and 15 years old and forty-five patients (6%) were between 1 day and 12 months old. The majority of the patients were enrolled into the treatment protocols of the Tokyo Children's Cancer Study Group (TCCSG) [6].
Cell separation and conventional hematological studies The heparinized bone marrow blood or peripheral blood were transferred to our laboratory for immunological studies from our medical department, affiliated hospitals of Keio University and the medical institutes contributing to the TCCSG study. The percentages of blasts in the samples were shown to exceed 60% in all 744 patients. Bone marrow or peripheral blood smears were stained by the May-Giemsa method. Conventional cytochemical stainings, such as myeloperoxidase, specific and non-specific esterase, Sudan-black, and periodic acid Schiff reaction were also performed. Mononuclear cells were separated by Hypaque-Ficoll density gradient centrifugation. After the cell separation, mononuclear cells were washed and suspended in phosphate buffer saline solution for immunological studies. In some cases, the separated cells were suspended in RPMI 1640 media supplemented with 20% fetal bovine serum and cultured without any mitogens or cytokines for 3 days in 5% CO, atmosphere (short-term culture). Rosette formation Rosette formation with sheep red blood cells (E-rosette) and erythrocyte-complementcomplex (EAC-rosette) were carried out by the method previously described [7]. Surface and cytoplasmic immunoglobulins Cell surface immunoglobulins (a, y. p, K and A chains) were determined by direct immunofluorescent staining using rabbit antia, y. p, K and h chains sera (Hoechst, Behringwerke, Germany). Cytoplasmic p chains were detected by the double peroxidase-antiperoxidase (PAP) method described elsewhere [81. Cell surface antigens From 1978 to 1980, indirect immunofluorescent staining was carried out for detecting cell surface antigens. The antibodies used in this period were well characterized rabbit anti-T cell, anti-CALLA and HLA-DR antisera
Actu
Puediutr Jpn
Inirnunological classific,ationof' A L L (89) 509
produced by the method previously described [9-111. From 1981, monoclonal antibodies were used by the micro-rosette method and flow cytometry. The micro-rosette method was developed by our laboratory [ 121 through the modification of previous methods [13, 141 for analyzing small numbers of cells. Briefly, the separated mononuclear cells were allowed to adhere on a Terasaki microtest plate (Falcon, #3034, USA) treated with po1y-Llysine. After the neutralization of positive charge of poly-L-lysine by fetal bovine serum, an adequate amount of monoclonal antibody was gently added, washed and ox red blood cells conjugated with goat-anti-mouse IgG antibody were added. After the reaction between the adhered mononuclear cells and the ox red blood cells, the microtest plate was placed upsidedown to remove non-reactive red cells from the bottom of the plate and the rosette forming cells were counted by light microscopy. The sensitivity of this method was compatible with those of indirect immunofluorescent staining. The clusters defined by monoclonal antibodies used in this study were CD6 (T6), CD2 (Tll), CD3 (T3), CD4 (T4), CD5 (TI), CD7 (Tp40), CD8 (T8) for T-lymphoid antigens, CDlO (J5, KOR-N34 [IS]), CD19 (B4), CD20 (BI), CD22 (Leul4) for B lymphoid antigens, C D l l (Leu15), CD13 (MY7, MCS2), CD14 (MY4), CD33 (MY9), CD36 (OKM5) for myelomonocytic antigens and HLA-DR antigen (KOR-la17 [15], 12, anti-HLA-DR). Heteroantisera (anti-T, anti-la, and anti-CALLA) produced in our laboratory were also utilized. In the late 1980's, CD41 (TP80), CD42 (AN51), CD56 (NKHI) and KOR-P77 [I61 (anti-platelet antigen) were added to the antibody panel and PFI, GTCSI were used for T-cell type ALL in recent studies. The reaction of the above antigens was defined as positive when antigen(s) were expressed on 30% or more cells determined by microrosette method or flow cytometry (Epics-CS, Coulter, USA). When blasts expressed both lymphoid and myelomonocytic antigens, double staining was performed.
Vol. 33 No. 4 August 1991
Cytoplasmic antigen From 1988, cytoplasmic antigens were determined by the double PAP method [8]. The antibodies used in this study were MY7 (CD13), T3 (CD3), P F l , GTCSI. The blasts from 277 patients with common ALL, 17 patients with unclassified ALL and some patients with biphenotypic leukemia were analyzed by using CD13. Cytoplasmic T-cell receptor antigens defined by T3, PFl, and GTCSl were examined in blasts from T-ALL and biphenotypic leukemias with T-cell marker( s). Diagnostic criteria The definition of ALL was based on FAB classification, conventional cytochemistry and the expression of lymphoid antigen(s) on the cell surface. Common ALL was diagnosed by the expression of CDIO, CD19, and/or CD22 and HLA-DR. Pre-B ALL was defined by the existence of cytoplasmic p chain in addition to the CDlO and CD19. The diagnostic markers of T-ALL and B-ALL were the expression of T-lymphoid antigens and E-rosette formation, and B-lymphoid antigens and surface immunoglobulins, respectively. Unclassified ALL was tentatively defined in terms of the expression of B-lymphoid antigens (CD19) and the loss of CDIO, T-lymphoid antigens and myelomonocytic antigens. In this study a diagnosis of biphenotypic leukemia was tentatively made when the expression of myeloid antigen(s) on the cell surface was proved in addition to lymphoid antigens. In this study patients with acute leukemia under 1 year were categorized as infant leukemia because of the different clinical features and poor prognosis. Ph' positive acute leukemia (Phi-AL) was diagnosed by the demonstration of chromosomal translocation t(9;22)(q34;qll) and bcr-abl chimeric mRNA. Cytogenetic study Cytogenetic analysis was performed in some patients from the late 1980's. G-banding and Q-banding methods were performed for chromosomal analysis in this study.
5 10 (90) N i r k u z ~ ~et~al. ~u
DNA and RNA analysis DNA and mRNA were extracted from the leukemic cells of patients with biphenotypic leukemias and infant leukemias. The rearrangement of T-cell receptor gene (Cp) and immunoglobulin heavy chain gene (JH) were studied by the standard Southern blotting method. Bcr-abl kimeric mRNA was defined by reverse transcriptase polimerase chain reaction (PCR) by using the following primers, major bcr exon 2 (b2) and 3 (b3), minor bcr exon 1 ( b l ) and abl exon 2 (a2), of which the characteristics have been described elsewhere [ 171.
Table I . The incidence of acute leukemia Patient number
522
ALL Common ALL T-ALL B-ALL Unclassified ALL
ANLL Biphenotypic leukemia Ph'-positive acute leukemia NK-leukemia Infant leukemia
404 (77.4%) 81 (15.5%) 21 (4.0%) 16 (3.1%) 139 29 8 1 45
A l 1.: acute Iymphoblautic leukemia. ANLL.: acute noii-l!mphocytic leubemia. N K : natural killer.
Results Incidence of ALL subtypes The overall incidence of subtypes of acute leukemias analyzed in this study is shown in Table 1. In 522 patients with ALL, 404 (77.4%). 81 (15.5%), 21 (4.0%) and 16 (3.1%) patients were diagnosed as common ALL, TALL. B-ALL and unclassified ALL, respectively. The incidence and prognosis of pre-B ALL could not be determined because of the small number in analyzed samples. Our preliminary study demonstrated that the 12 out of 44 common ALL (27.3%) which had cytoplasmic /.I chain, were not different from that of common ALL in an event free survival (EFS). Ph'-AL could not be differentiated from common ALL before the 1986's. In all seventeen patients with unclassified ALL, only CD19 antigen was expressed and other lymphoid antigens could not be detected. The ALL'ANLL ratio (4:I ) does not reflect the true incidence in the TCCSG study, since the blood samples from patients with ANLL were not transferred before the middle of the 1980's. There was one patient with blasts expressing natural killer activity and CD56 (NKH-I). The marker profiles and clinical features have been described elsewhere [ 181. Common ALL Common ALL is the most common subtype among childhood acute leukemias. The majority of common ALL blasts expressed
B-lineage antigens (CD10, CD19, CD22 and HLA-DR). In this study the expression of cell surface CD13 antigen was examined in 277 patients with common ALL. In 10 out of 277 cases, blasts expressed CD13. These 10 cases were categorized as biphenotypic leukemia. Cytoplasmic CD13 (cCD13) was examined in 59 newly diagnosed common ALL cases in which the expression of surface CD13 was negative. In 11 out of 59 patients (18.6%), cCD13 antigen was demonstrated. After short term culture without any mitogens or cytokines, surface CD13 appeared in all 11 cases. With the extensive studies including karyotype analysis and mRNA by PCR, five out of these 1 I patients were categorized as Ph'-AL based on the existence of t(9;22) (q34; q l l ) translocation and bcr-abl kimeric mRNA. T-ALL In an earlier study, heteroantiserum and Erosette formation were used for the diagnosis of T-ALL, since the monoclonal antibodies were not available. Thus these cases were reexamined by monoclonal antibodies using recovered leukemic cells which were stored in liquid nitrogen and the diagnosis was reconfirmed. Fig. I shows the incidence of cell surface and cytoplasmic antigens in 42 patients with T-ALL. In all cases, CD7 and cytoplasmic CD3 (cCD3) were expressed. Table 2 shows the proposal of the subclassification of these 42 patients based on the
A d a Paediatr Jprz
Inimunological classi$carion of' A L L (91) 51 I
sCD sCD sCD sCD sCD sCD
0
20
40
60
80
100 ( 0 4
Positive Cases Fig. I : Incidence of surface and cytoplasmic T-antigens in T-ALL. CDI, 3. 4, 8, 5 , 2. 7 were defined on the cell surface by micro-rosette method. T-ALL: Forty-two cases were well characterired for cell surface and cytoplasmic antigens. SFI : recogniie TcRP chain.
expression of T-cell receptor (TcR) antigens. The most common pattern was cbTCS1-, cPFI+, sCD3- subtype. The clinical significance of this classification is now under investigation. In this study, further classifications based on other T-cell differentiation antigens were tried (data not shown). The pattern of antigen expression was heterogeneous and it was impossible to divide T-ALLs into subgroups as previously described [19, 201. B-ALL The blasts in B-ALL expressed cell surface immunoglobulin, CDIO, CD19, CD20, and CD22. The most common immunoglobulin subtype was IgM @). The morphology of these were FAB-L3 (18 cases) and L1 (3 cases). There were some cases indistinguishable from Burkitt's lymphoma at stage IV. B-ALL had a poor prognosis even if intensive chemotherapy was employed. The blasts with LI morphology (3 cases) displayed a small amount of Y, chain on the cell surface [21]. Biphenotypic leukemias Although the definition, nomenclature and diagnostic criteria of biphenotypic leukemia
Vol. 33 No. 4 August 1991
Table 2. Subclassification of T-ALL based on T-cell receptor antigens d T C S I i cBFl i sCD3
-i-I+I-i-I+--I+ or - 1 -
Incidence (%)
8 (19%) 1(2%) 23 (55%) 10 (24%)
T-AI.l.: Forty-two cases were well characteriied for cell surface and cytoplasmic antigens. c6TCS I : cytoplasmic TcK 6 chain recognized by GTCSI. cPFI: cytoplasmic TcR chain recogni7ed by PFI. sCD3: cell surface CD3 antigen recogniied by T3.
have not been established, the diagnosis of biphenotypic leukemia was tentatively made in this study when one of the following criteria was fulfilled: (1) the coexpression of both lymphoid and myelomocytic antigen(s) on the same leukemic cells (biphenotypic), (2) coexistence of the cells expressing different cell lineage characteristics (lymphoid and myelomonocytic) and the cells were proved to have the same clonal origin, and (3) the cell lineage changed during the course of the disease (lineage switch). Twenty-nine out of 744 patients were diagnosed as biphenotypic leukemia. These patients were divided into 4
5 I2 (92) Naliu:urru Pt at.
myeloid antigen(s) were confirmed by double staining on flow cytometry. The antigenic markers of these cases are summarized in Tables 3 to 7. In 11 patients with common ALL (group I), the leukemic cells expressed CD13 (lo/ 11). In one case (patient #ll), CD36 was expressed in addition to CDlO, CD19 and HLA-DR. In five patients with unclassified ALL (group 2), heterogeneity of
groups as follows: ( 1 ) common ALL with positive myelomonocytic antigen(s) expression, (2) unclassified ALL with positive myelomonocytic antigen(s) expression, (3) ANLL (Ml) with positive B- lymphoid antigen(s) expression and (4) acute leukemia with positive T- and myeloid antigen(s) expression (T/M hybrid leukemia). In the majority of these cases, biphenotypic expression of lymphoid and
Table 3. Group I , common ALL with positive myelomonocytic antigen(s) expression Patient no.
Age (Yr)
WBC
1
15600
'1.
50800 3800 I600 7400 8800 3900 5900 32600 I300 139900
I. 2. 3. 4. 5. 6.
4 4
7. 8. 9. 10. II .
8 9 II 12 14
S
5
(:uU
HLA-DR (12)
CDlO (JS)
CD19
(B4)
CD20 (BI)
86 93 88 96 43 92 85 97 86 100 87
87 98 64 96 98 92 83 80 83 96 80
nt
nt
98 83 9s nt 96 84 nt 82 nt 91
17 43 53 nt
