CD34-positive acute promyelocytic leukemia is associated with leukocytosis,...

8
CD34-Positive Acute Promyelocytic Leukemia Is Associated With Leukocytosis, Microgranular/ Hypogranular Morphology, Expression of CD2 and bcr3 Isoform R. Foley, 1,2,3 P. Soamboonsrup, 1 R.F. Carter, 2 A. Benger, 1,3 R. Meyer, 1,3 I. Walker, 3 Y. Wan, 2 W. Patterson, 1 A. Orzel, 1 L. Sunisloe, 1 B. Leber, 1,2,3 and P.B. Neame 1,2 * 1 Department of Laboratory Medicine, Hamilton Health Sciences Corporation, Hamilton, Canada 2 Department of Pathology and Molecular Medicine, McMaster University, Hamilton, Canada 3 Department of Medicine, Hamilton Health Sciences Corporation and McMaster University, Hamilton, Canada Acute promyelocytic leukemia (APL) has a favorable prognosis. Current therapy includes chemotherapy used in combination with all-trans-retinoic acid (ATRA). Although the dif- ferentiating effects of ATRA on promyelocytes have been well established, in vitro stud- ies have shown that less-differentiated APL blasts (CD34 + ) demonstrate a variable re- sponsiveness to ATRA. To assess the clinical relevance of this finding, we analyzed a cohort of 38 patients with t(15;17) and/or PML-RARa APL to determine the incidence and laboratory features of CD34 + APL. Thirty-two percent (12/38) of cases were CD34 + . There was a difference in WBC at presentation between CD34 + and CD34 - cases (34.6 ± 9.2, mean ± standard error vs. 5.4 ± 2.0, P = 0.009). Patients with CD34 + APL demonstrated a micro/hypogranular phenotype (75%) (P = 0.001), co-expression of CD2 + (83%) (P = 0.001), and the bcr3 isoform (100%) (P = 0.017). In contrast, CD34 - cases demonstrated hyper- granular morphology (65%), CD2 + (15%), and the bcr1 isoform (50%). A high presenting WBC count (10 × 10 9 /L) was associated with an inferior overall survival (Log rank = 0.0047). Patients with CD34 + APL demonstrated an incidence of early mortality of 50%. Despite a marked correlation between CD34 positivity and increased WBC count, overall survival of CD34 + and CD34 - cases did not differ significantly in our small cohort. Im- munophenotypic analysis for CD34 expression should be included in future large APL trials to determine if detection of CD34 + blasts represents an independent adverse prog- nostic factor. Am. J. Hematol. 67:34–41, 2001. © 2001 Wiley-Liss, Inc. Key words: APL; PML-RARa; CD34; CD2; all-trans-retinoic acid INTRODUCTION Acute promyelocytic leukemia (APL) defined by cy- togenetic t(15;17) or molecular (PML-RARa) analysis has a relatively favorable prognosis [1–3]. Several ran- domized trials have demonstrated that the administration of ATRA in combination with chemotherapy improves overall survival [1,4,5]. Despite this success, up to one- third of patients with t(15;17) APL relapse after attaining complete remission [2,3,5,6]. Identification of adverse prognostic markers in APL is important and potentially useful in the development of treatment strategies for pa- tients with high-risk disease. Factors associated with an inferior prognosis in t(15;17) APL include older age [3,5,7], increased WBC count at presentation [1,5,7,8], and the detection of the PML-RARa translocation prod- uct by RT-PCR during consolidation or following completion of therapy [2–4,6]. Other factors such as ad- ditional chromosomal abnormalities [1,9] or molecular PML-RARa subtype [10–12] do not appear to predict clinical outcome [1,9–12]. Immunophenotypic markers have also been examined Contract grant sponsor: Leukemia Research Foundation of Canada; Contract grant number: 77681. *Correspondence to: Dr. P.B. Neame, Hamilton Regional Laboratory Medicine Program, Henderson General Hospital Site, 711 Concession Street, Hamilton, Ontario, L8V 1C3, Canada. E-mail: [email protected] Received for publication 26 June 2000; Accepted 15 November 2000 American Journal of Hematology 67:34–41 (2001) © 2001 Wiley-Liss, Inc.

Transcript of CD34-positive acute promyelocytic leukemia is associated with leukocytosis,...

Page 1: CD34-positive acute promyelocytic leukemia is associated with leukocytosis, microgranular/hypogranular morphology, expression of CD2 and bcr3 isoform

CD34-Positive Acute Promyelocytic Leukemia IsAssociated With Leukocytosis, Microgranular/

Hypogranular Morphology, Expression of CD2 andbcr3 Isoform

R. Foley, 1,2,3 P. Soamboonsrup, 1 R.F. Carter, 2 A. Benger, 1,3 R. Meyer,1,3 I. Walker, 3 Y. Wan,2

W. Patterson, 1 A. Orzel, 1 L. Sunisloe, 1 B. Leber, 1,2,3 and P.B. Neame 1,2*1Department of Laboratory Medicine, Hamilton Health Sciences Corporation, Hamilton, Canada

2Department of Pathology and Molecular Medicine, McMaster University, Hamilton, Canada3Department of Medicine, Hamilton Health Sciences Corporation and McMaster University, Hamilton, Canada

Acute promyelocytic leukemia (APL) has a favorable prognosis. Current therapy includeschemotherapy used in combination with all-trans -retinoic acid (ATRA). Although the dif-ferentiating effects of ATRA on promyelocytes have been well established, in vitro stud-ies have shown that less-differentiated APL blasts (CD34 +) demonstrate a variable re-sponsiveness to ATRA. To assess the clinical relevance of this finding, we analyzed acohort of 38 patients with t(15;17) and/or PML-RAR a APL to determine the incidence andlaboratory features of CD34 + APL. Thirty-two percent (12/38) of cases were CD34 +. Therewas a difference in WBC at presentation between CD34 + and CD34− cases (34.6 ± 9.2,mean ± standard error vs. 5.4 ± 2.0, P = 0.009). Patients with CD34 + APL demonstrated amicro/hypogranular phenotype (75%) ( P = 0.001), co-expression of CD2 + (83%) (P = 0.001),and the bcr3 isoform (100%) ( P = 0.017). In contrast, CD34 − cases demonstrated hyper-granular morphology (65%), CD2 + (15%), and the bcr1 isoform (50%). A high presentingWBC count ( ≥10 × 109/L) was associated with an inferior overall survival (Log rank =0.0047). Patients with CD34 + APL demonstrated an incidence of early mortality of 50%.Despite a marked correlation between CD34 positivity and increased WBC count, overallsurvival of CD34 + and CD34− cases did not differ significantly in our small cohort. Im-munophenotypic analysis for CD34 expression should be included in future large APLtrials to determine if detection of CD34 + blasts represents an independent adverse prog-nostic factor. Am. J. Hematol. 67:34–41, 2001. © 2001 Wiley-Liss, Inc.

Key words: APL; PML-RAR a; CD34; CD2; all-trans- retinoic acid

INTRODUCTION

Acute promyelocytic leukemia (APL) defined by cy-togenetic t(15;17) or molecular (PML-RARa) analysishas a relatively favorable prognosis [1–3]. Several ran-domized trials have demonstrated that the administrationof ATRA in combination with chemotherapy improvesoverall survival [1,4,5]. Despite this success, up to one-third of patients with t(15;17) APL relapse after attainingcomplete remission [2,3,5,6]. Identification of adverseprognostic markers in APL is important and potentiallyuseful in the development of treatment strategies for pa-tients with high-risk disease. Factors associated with aninferior prognosis in t(15;17) APL include older age[3,5,7], increased WBC count at presentation [1,5,7,8],and the detection of the PML-RARa translocation prod-

uct by RT-PCR during consolidation or followingcompletion of therapy [2–4,6]. Other factors such as ad-ditional chromosomal abnormalities [1,9] or molecularPML-RARa subtype [10–12] do not appear to predictclinical outcome [1,9–12].

Immunophenotypic markers have also been examined

Contract grant sponsor: Leukemia Research Foundation of Canada;Contract grant number: 77681.

*Correspondence to: Dr. P.B. Neame, Hamilton Regional LaboratoryMedicine Program, Henderson General Hospital Site, 711 ConcessionStreet, Hamilton, Ontario, L8V 1C3, Canada.E-mail: [email protected]

Received for publication 26 June 2000; Accepted 15 November 2000

American Journal of Hematology 67:34–41 (2001)

© 2001 Wiley-Liss, Inc.

Page 2: CD34-positive acute promyelocytic leukemia is associated with leukocytosis, microgranular/hypogranular morphology, expression of CD2 and bcr3 isoform

in APL. Typical surface markers include CD13, CD33,and CD9 [12–14]. Atypical markers, including CD2 andCD34, have been reported [12,15]. Although the clinicalrelevance of these immunophenotypic markers is notclearly known, several groups have shown that expres-sion of CD2 is associated with microgranular morphol-ogy (M3v) [16,17] and the bcr3 PML-RARa isoform[12,18]. In addition, the neural cell adhesion moleculeCD56 has recently been shown to predict poor outcomein patients treated with ATRA and simultaneous chemo-therapy[19].

Until recently, the detection of CD34 in t(15;17) APLwas considered infrequent [14]; however, several reportshave now shown that expression of CD34 occurs in asignificant proportion of newly diagnosed t(15;17) APLpatients [12,20–22]. Although CD34 positivity has beenassociated with a lower rate of complete remission incombined subtypes of acute myeloid leukemia (AML)[23–25], the significance of CD34 expression in t(15;17)/PML-RARa APL is not known. It is possible that detec-tion of the CD34 antigen may signify an immature formof APL. Because current standard therapy involves theadministration of ATRA for induction and/or mainte-nance, it will be useful to know how these immature cellsrespond to differentiating agents such as ATRA [26,27].Recent in vitro studies have demonstrated that culturedAPL stem cells, as opposed to promyelocytes, demon-strate a variable responsiveness to ATRA [27]. Similar invitro findings have been shown for selected CD34+ pri-mary APL cells [28]. Whether the presence of CD34+

APL leukemic cells at diagnosis results in an inferioroutcome following ATRA and current cytotoxic therapyis unknown.

To characterize the laboratory and clinical features ofCD34+ APL and determine the proportion of positivecases, we analyzed 38 consecutive patients with t(15;17)and/or PML-RARa APL. A diagnosis of APL was es-tablished by cytogenetics, FISH, Southern blot, or RT-PCR. Results of this analysis demonstrate CD34+

t(15;17)/PML-RARa APL occurs in up to one-third ofAPL cases and is associated with an immature micro/hypogranular phenotype. In addition, expression ofCD34 was associated with co-expression of CD2 and thepresence of the PML-RARa bcr3 (short) breakpoint.There was a highly significant correlation between CD34expression and increased WBC count, an established ad-verse marker in APL that is associated with an increasedrate of relapse [1,5]. Inclusion of CD34 expression incurrent large multicenter clinical trials may help to de-termine if the detection of a subpopulation of CD34+

APL blasts is an independent adverse prognostic marker.

MATERIALS AND METHODSPatients

Peripheral blood and bone marrow samples from 38consecutive cases of newly diagnosed APL were ana-

lyzed between January 1988 and January 2000 at theMalignant Hematology Diagnostic Unit (HamiltonHealth Sciences Corporation, Hamilton, Canada). Allcases were confirmed as t(15;17) by karyotyping or FISHand/or as PML-RARa by RT-PCR or Southern blot.Clinical variables analyzed included (i) age, (ii) treat-ment protocol, and (iii) survival. Laboratory analysis in-cluded (i) presenting WBC count, (ii) immunophenotyp-ing, and (iii) evaluation of disseminated intravascularcoagulopathy (DIC)/fibrinolysis based on a fibrinogenlevel of <1.0 g/l without liver disease and a positiveprotamine sulfate or D-dimer test. Standard microscopicand cytochemical evaluations were used to classify mor-phological subtypes using standard FAB criteria [29].Morphological classification included M3, M3 variant(M3v) [30], or mixed (M3-M3v). Other cases were clas-sified as M1 or M2-like APL [15] or as a hypogranularbasophilic variant [15,31]. Wright-Giemsa stained bonemarrow and peripheral blood preparations of all caseswere examined by two pairs of morphologists. Each pairexamined the samples and reached a morphological clas-sification. Any case where there was disagreement wasre-examined by the senior members of the pairs, and aconsensus diagnosis was made. All examinations wereperformed without knowledge of specific clinical or im-munophenotypic data.

Immunophenotypic Analysis

Immunophenotypic analysis was performed on bonemarrow cells in all but three cases (cases 1, 3, and 6)where peripheral blood was used. A panel of monoclonalantibodies was used to establish the surface antigen pro-file of APL cases studied. These included CD33, CD13,CD14, CD2, CD19, CD11b, HLA-DR, and CD117 fromBeckman-Coulter Corp (Hialeah, FL), CD34 [clone 581-Immunotech/Beckman-Coulter or HPCA2-Becton Dick-inson (San Jose, CA)] and CD15, CD5, and CD11c (Bec-ton Dickinson). The percentage of leukemia cells stainedfor a given MoAb was determined by selecting blasts andpromyelocytes on the basis of forward and 90° light-scatter signals. A positive reaction was defined when$20% of APL cells expressed more fluorescence inten-sity than control cells. Expression of CD2 in the absenceof other pan-T antibodies (CD5 and/or CD3) was used toexclude contamination with T lymphocytes. In addition,double staining using fluorescein isothiocyanate (FITC)and phycoerythrin (PE) was performed in selected casesto confirm CD2 antigen (FITC) co-expression, used incombination with other PE conjugated antibodies, e.g.,CD33 or CD34. An anti CD34 monoclonal antibody wasutilized to detect an underlying CD34+ subpopulation.We and other investigators [8,12,20] have used a 10%cutoff to quantify the presence of a subpopulation ofCD34+ cells as opposed to 20% used for disease charac-terization.

CD34-Positive APL 35

Page 3: CD34-positive acute promyelocytic leukemia is associated with leukocytosis, microgranular/hypogranular morphology, expression of CD2 and bcr3 isoform

Cytogenetic and FISH Analysis

Cytogenetic analysis was performed on marrowsamples taken at the time of initial diagnosis. Routineharvesting methods were used to prepare slides for G-banding, digital imaging, and computer-assisted karyo-typing (PSI Instrument, League City, TX). Metaphaseswere analyzed directly or after 24 hr of culture in RPMI1640 (Gibco, New York, NY) at 37°C with 5% CO2. ForFISH analysis, metaphase and interphase cells were sub-jected to two-color fluorescence analysis after hybridiza-tion using a 15;17 probe set (P5119- P/B, Oncor, Gai-thersburg, MD). Probe sequences labeled withrhodamine (17q21 locus) or fluorescein (15q22 locus)were visualized by fluorescence microscopy using a dualband-pass filter (Zeiss, Toronto, Ontario). Southern blotanalysis was performed by methods previously described[15].

RT-PCR

Total RNA was extracted from bone marrow mono-nuclear cells using TRIzol reagent (Life Sciences Tech-nologies, Gaithersburg, MD) according to the manufac-turer’s specifications. Reverse transcription and PCR ofthe PML-RARa break/fusion product were completedusing a standard protocol as described by Gallagher et al.[32]. Classification of the specific PML-RARa isoform[(bcr1) long, (bcr2) variable, and (bcr3) short] was evalu-ated in 19 of 38 cases.

Statistical Analysis

An SPSS statistical package (version 8) was used toanalyze all laboratory and clinical data. Patient data werecoded and stored in a data base. Statistical measures in-cluding t-test for equality of means, Chi-square (Fisher’s

exact test), and Kaplan–Meier analysis were performeddirectly from this data base.

RESULTSAssociation Between CD34 Positivity andMorphological Subtype

Clinical and laboratory data for both CD34+ andCD34− patients is summarized in Tables I and II. Ex-pression of CD34 ($10% marrow mononuclear cells)was found in 12/38 cases (32%). The range of CD34positivity varied considerably from 11% to 95%. Nine of12 patients with CD34+ APL demonstrated a micro/hypogranular phenotype (75%). The frequency of micro/hypogranular morphology was statistically higher inCD34+ cases when compared to CD34− cases (Fisher’sexact test,P 4 0.001). Within the CD34+ subgroup, fourcases demonstrated FAB variant (M3v) morphology andfour cases showed basophilic variant morphology. Onecase demonstrated M1-like blast morphology [33]. Onlythree (25%) CD34+ patients demonstrated typical FABM3 hypergranular morphology. Conversely, 17 of 26(65%) CD34− cases revealed hypergranular morphology(FAB M3).

Association Between CD34 Positivity andElevated WBC Count at Presentation

Patients with CD34+ APL demonstrated significantleukocytosis at presentation (WBC4 34.6 ± 9.2 × 109/L,mean ± SEM) (Fig. 1). CD34− patients presented with asignificantly lower WBC (5.4 ± 2.1 × 109/L, P 4 0.009).Leukocytosis at presentation was also noted in patientsexpressing CD2 ($20%) (Fig. 1). Chi-square analysiswas used to assess the relationship between CD34 posi-tivity and high presenting WBC count. In this analysis a

TABLE I. Clinical and Laboratory Data of CD34 + APL Cases*

Caseno.a

Patientage/sex Morphology WBC DIC

CD34+

(%)CD2+

(%)CD19+

(%) t(15;17)PML/RARa

isoform Treatment

1 35F FAB variant30 35.0 + 28 74 1 +b Chemo2 68M FAB variant 27.0 + 78 86 1 +b Short/bcr3 Cryo/plts3 33M FAB variant 46.2 + 14 98 N/A N/A Short/bcr3 Cryo/plts4 32M Basophilic variant15 23.9 + 57 87 1 + ATRA/chemo/ABMT5 58F Basophilic variant 16.7 + 11 94 94 + ATRA/chemo6 65M Basophilic variant 8.6 + 52 58 41 + ATRA/chemo7 46M M1-like APL15,33 77.7 + 95 96 2 −c Short/bcr3 ATRA/chemo8 73F FAB variant 0.8 − 16 83 1 − Short/bcr3 ATRA/chemo9 49M Hypergranular29 0.5 + 11 1 1 + ATRA/chemo

10 33M Hypergranular 31.7 + 12 8 1 −c Short/bcr3 ATRA/chemo11 57M Hypergranular 37.0 + 15 19 0 + Short/bcr3 ATRA/chemo12 23F Basophilic variant 107.8 + 68 43 80 + Short/bcr3 ATRA/chemo

*Chemo4 chemotherapy; Mitoxantrone and Cytarabine, high-dose Cytarabine (see Materials and methods); Cryo/plts4 transfusion with cryoprecipitateand platelets only; ABMT4 autologous bone marrow transplant; DIC4 disseminated intravascular coagulopathy/fibrinolysis; N/A4 not available.aFor descriptive purposes only, cases have been labeled from 1 to 12.bAdditional chromosomal abnormality.cChromosomal abnormality other than t(15;17).

36 Foley et al.

Page 4: CD34-positive acute promyelocytic leukemia is associated with leukocytosis, microgranular/hypogranular morphology, expression of CD2 and bcr3 isoform

high WBC count, defined as$10 × 109/L correlated withCD34 positivity ($10%) (Fisher’s exact test,P 4 0.02).A similar association was found between CD2 and highWBC count (P 4 0.02). A modest correlation betweenrange of CD34 expression and range of CD2 expression(CD2+/CD34+) was also found (R 4 0.51). Chi-squaredanalysis evaluating the association between CD34 posi-tivity and CD2 positivity was highly significant (Fisher’sexact test,P 4 0.001). Cases expressing high-level CD2without CD34 or cases positive for CD34 without CD2were noted. Patients positive for both CD34 and CD2 (N4 9) had a mean WBC count of 38.2 ± 11.5 × 109/L.Patients positive for CD34 and negative for CD2 (N 4 3)also demonstrated a high WBC count (23.1 ± 11.4 ×109/L). Cases positive for CD2 but negative for CD34demonstrated a mean WBC of 4.5 ± 1.8 × 109/L (N 4 4).

Molecular and Karyotypic Analysis

Results of molecular and karyotypic studies are sum-marized in Tables I and II. RT-PCR PML-RARa isoformanalysis was completed in 19 of 38 cases. Patients withCD34+ APL demonstrated the short (bcr3) form of PML-

TABLE II. Clinical and Laboratory Data on CD34 − APL Cases*

Caseno.a

Patientage/sex Morphology WBC DIC

CD34+

(%)CD2+

(%)CD19+

(%) t(15;17)PML/RARa

isoform Treatment

13 27F Hypergranular29 1.9 + 0 2 3 + None14 66M Basophilic variant15 8.0 + 2 55 4 + Cyro/plts15 40F Hypergranular 1.3 + 2 1 5 +b ATRA/chemo/AlloBMT16 47M Hypergranular 2.0 + 1 1 15 +b Cryo/plts17 70F Hypergranular 0.5 − 4 0 0 + None18 7F Hypergranular 1.2 + 0 0 0 + ATRA/chemo/AlloBMT19 43F Hypergranular 4.6 − 1 1 1 + Long/bcr1 ATRA/chemo20 35M Hypergranular 53.8 + 0 3 1 + ATRA21 62M Hypergranular 0.8 − 1 3 1 + Long/bcr1 ATRA/chemo22 76F Hypergranular 0.6 + 1 10 1 + Short/bcr3 ATRA23 62M Hypergranular 2.2 + 3 1 1 +b Long/bcr1 ATRA/chemo24 18F Hypergranular 0.8 − 6 6 2 + Long/bcr1 ATRA/chemo25 30F Hypergranular 7.1 − 4 2 1 −c Long/bcr1 ATRA/chemo26 38M Hypergranular 1.1 + 0 0 1 + Long/bcr1 ATRA/chemo27 76F Hypergranular 1.0 + 1 1 1 + Short/bcr3 ATRA alone28 64F Hypergranular 1.0 N/A 1 1 2 + N/A29 35M M2-like15 11.2 + 1 1 0 + ATRA/chemo30 54F M2-like 3.8 + 0 7 1 FISH ATRA/chemo31 54M M2-like 13.1 − 3 3 1 +b Cryo/plts32 60F M1-like15 5.1 N/A 1 2 3 −c *d None33 34F Mixed 1.2 + 0 1 1 − Variable/bcr2 ATRA/chemo34 24F FAB variant30 0.9 − 0 99 1 + Short/bcr3 ATRA/chemo35 54M Hypergranular 1.7 + 5 3 2 +b Short/bcr3 ATRA/chemo36 45M Hypergranular 1.5 + 5 5 3 − Short/bcr3 ATRA/chemo37 84F Basophilic variant 2.1 + 1 78 0 + ATRA alone38 71F Basophilic variant 7.1 + 9 33 2 + ATRA/chemo

*Chemo4 chemotherapy; Mitoxantrone and Cytarabine, high-dose Cytarabine (see Materials and methods); Cryo/plts4 transfusion with cryoprecipitateand platelets only; AlloBMT4 allogeneic bone marrow transplant; DIC4 disseminated intravascular coagulopathy/fibrinolysis; N/A4 not available.aFor descriptive purposes only, cases have been labeled from 13 to 38.bAdditional chromosomal abnormality.cChromosomal abnormality other than t(15;17).dSouthern blot PML-RARa.

Fig. 1. Mean WBC count at presentation in APL patientspositive for CD34 antigen expression APL cases were clas-sified as CD34 + or CD34− and as CD34 +/CD2+, CD34+/CD2−,and CD34−/CD2+. There was a difference in WBC count atdiagnosis between CD34 + (34.6 ± 9.2) and CD34 − (5.4 ± 2.0)cases ( P = 0.009).

CD34-Positive APL 37

Page 5: CD34-positive acute promyelocytic leukemia is associated with leukocytosis, microgranular/hypogranular morphology, expression of CD2 and bcr3 isoform

RARa in all (7/7) cases analyzed. Demonstration of bcr3was significantly higher (P 4 0.017) in the CD34+ APLsubgroup as compared to the CD34− subgroup where6/12 (50%) demonstrated the long (bcr1) form and 5/12(42%) were found to have the short form (bcr3). Onecase demonstrated the bcr2 isoform. The t(15;17) trans-location was demonstrated by standard karyotyping in29/38 (76%) cases. Additional chromosomal abnormali-ties were noted in both CD34+ [4/12, (33%)] and CD34−

[7/26, (27%)] subgroups.

Clinical Outcome of CD34 + Cases

Thirty eight patients (19 male and 19 female, medianage 4 46.5) with t(15;17) or PML-RARa APL wereanalyzed. Seventy-four percent of cases (28/38) receivedATRA either in induction and/or for maintenance over 1year. Twenty-four (63%) patients received ATRA withchemotherapy. Cytotoxic therapy included Mitoxantrone(12 mg/m2 continuously × 3 days) and cytarabine (100mg/m2 continuously × 7 days) for induction and firstconsolidation. A second consolidation included high-dose cytarabine (2 g/m2 × 8 doses). Two patients re-ceived ATRA alone (cases 27 and 37), and two patientsstarted ATRA (cases 20 and 22) but died prior to initi-ating chemotherapy. Five patients died early after receiv-ing supportive cryoprecipitate and platelet transfusionsonly (cases 2, 3, 14, 16, and 31). At the time of analysis16/38 (42%) patients were alive with a median follow-upof 32.4 months.

Patients with CD34+ APL demonstrated a 50% inci-dence of early death (cases 1–3, 5, 6, and 11). In thesecases, 4/6 patients died of intracerebral hemorrhage(cases 1, 2, 3, and 11) and two patients died of sepsis(cases 5 and 6). However, the incidence of early death inCD34+ and CD34− groups was not significantly different(P 4 0.3). The overall survival of patients with a highWBC count ($10 × 109/L) at presentation was inferior topatients with a presenting WBC count <10 × 109/L (Logrank P 4 0.0047; Fig. 2). The survival of CD34+ APLpatients was compared to CD34− patients and despite amarked association with high WBC count, expression ofCD34 did not correlate with a statistically different sur-vival outcome (Log rankP 4 0.17; Fig. 3).

DISCUSSION

APL is usually identified by characteristic morphology(M3 or M3v) and an immunophenotypic profile that in-cludes CD13+, CD33+, CD9+, HLA-DR−, and CD14−

[12–15]. Recent utilization of molecular testing has led tothe identification of PML-RARa positive variants withatypical morphology [15,33]. Although expression ofCD34 was initially considered uncommon [14], recentstudies have shown that CD34 positivity occurs in about20–30% of newly diagnosed cases [12,20–22]. The sig-

nificance of CD34 expression in APL is unknown butlikely identifies an immature form of APL. Our findingsassociate CD34 positivity with leukocytosis, micro/hypogranular morphology, expression of CD2 and bcr3isoform. Despite some overlap, CD34 expression sepa-rated cases of t(15;17)/PML-RARa APL into two broadcategories, including (i) CD34−, low WBC count, typicalhypergranular morphology, CD2−, as compared to (ii)CD34+, high WBC count, micro/hypogranular morphol-ogy, and co-expression of CD2+.

The most significant characteristic of CD34+ APL was

Fig. 2. Survival of APL patients based on WBC count atpresentation. Kaplan–Meier survival analysis was used tocompare the overall survival based on presenting WBCcount. Patients with a WBC count greater than or equal to 10× 109/L. demonstrated an inferior outcome (Log rank P =0.0047).

Fig. 3. Overall survival of APL patients based on the ex-pression of CD34 +. Immunophenotypic analysis of APLcells was used to classify patients as CD34 + or CD34−. Over-all survival using Kaplan–Meier analysis is shown. Althoughthere was a trend towards poor outcome in the CD34 + sub-group, overall survival was not significantly different (Logrank P = 0.17).

38 Foley et al.

Page 6: CD34-positive acute promyelocytic leukemia is associated with leukocytosis, microgranular/hypogranular morphology, expression of CD2 and bcr3 isoform

an association with a high presenting WBC count. Sev-eral studies have consistently demonstrated that a pre-senting WBC count (greater than 5 × 109/L or 10 ×109/L) correlates with an inferior clinical outcome inAPL [1,4,5,7,8]. Feneaux et al. [5] have shown that olderage (P 4 0.006) and higher WBC count upon inclusion(P 4 0.02) are predictive of increased risk of early death.Patients with a high WBC count have a greater risk ofearly death due to hemorrhage [34,35]. This may be re-lated to DIC, fibrinolysis, nonspecific proteolysis, andthrombocytopenia [36,37]. However, early death is notthe only factor resulting in a poor outcome in APL pa-tients with a high WBC count. Both the recent MRC andEuropean trials [1,5] have shown that increased WBCcount is also associated with an increased risk of relapse.Moreover, Burnett et al. [1] have shown that patientswith a high presenting WBC count demonstrate an at-tenuated response to ATRA. Specifically, the investiga-tors found that an extended course of ATRA, as opposedto a short induction, results in an overall significant re-duction in relapse (20% vs. 36%,P 4 0.04) and im-proved overall survival from diagnosis in patients withAPL (71% vs. 52% at 4 years,P 4 0.005). However, anybenefits attributable to ATRA were seen only in caseswith a low WBC count and not in patients with a highpresenting WBC count ($10 × 109/L) [1].

It is important to define or characterize the biologicalfeatures of APL demonstrating a high WBC count. Fac-tors previously associated with a high WBC count in-clude expression of CD2, M3v morphology, and theshort form of PML-RARa [10,16,17,38]. The outcomeof patients with M3v is considered to be inferior to caseswith typical M3 morphology [2]. Guglielmi et al. [12]recently assessed the expression of CD2 and CD34 in197 patients with APL confirmed by molecular analysis.The frequencies of these markers were 28% and 23%,respectively. CD2 expression was associated with M3vmorphology, bcr3, and CD34 positivity. In addition,CD19 was found to correlate with the WBC count [12].

All CD34-positive cases in our analysis demonstratedthe short bcr3 isoform of PML-RARa by RT-PCR analy-sis. The clinical significance of bcr3 as a prognosticmarker is controversial. Several investigators, includingGallagher et al. [10], have reported that the presence ofthe PML-RARa translocation product does not predictclinical outcome [10–12]; however, others have sug-gested that bcr3 is associated with a inferior outcome[39–41]. Reasons for these discrepant findings may berelated to different treatment strategies used by variousgroups and the relatively small sample size in a numberof studies [1].

The combination of ATRA and cytotoxic chemo-therapy has improved the outcome of patients diagnosedwith t(15;17)/PML-RARa APL with overall survivalcurrently in the range of 70% [1,4,5,6]. The benefit of

ATRA in the setting of a high WBC count is less clear[1]. In interpreting resistance to ATRA, most emphasishas been placed on acquired mechanisms related to thepharmacokinetics of ATRA [42–44]; or to the develop-ment of molecular mutations in the RARa portion of thefusion gene [45–48]. Less emphasis has been placed onthe possibility of primary intrinsic mechanisms leadingto a poor prognosis. One such factor may include thestate of maturation as defined by immunophenotypicmarkers. In these cases, immature CD34+ APL blastsmay be less responsive to differentiating agents. In vitrostudies support this contention [27,28,49,50]. ThoughATRA is highly effective in inducing morphologicalmaturation of leukemic promyelocytes, the effect ofATRA on APL blast stem cells is variable [27]. Further-more, CD34-selected APL cells have a delayed responseto ATRA and retain a blastic phenotype, suggesting thatCD34+ blast cells may persist in patients followingATRA therapy [28]. Thus, the expression of CD34 onAPL cells at diagnosis may identify patients at a higherrisk for relapse. Different treatment regimens may needto be considered for these immature/high WBC variants.

Most large clinical trials have not included CD34 im-munophenotypic data in their published reports[1,4,5,7,10], and although a trend toward a poor outcomewas noted in our study, we were unable to demonstrate astatistical difference in the survival outcome of CD34-positive and CD34-negative cases in our small cohort.These considerations suggest that future large trials inAPL should include CD34 measurement to determine ifthe presence of a proportion of CD34+ APL blasts atpresentation is an independent adverse prognosticmarker.

ACKNOWLEDGMENTS

We thank the members of the Malignant DiagnosticUnit at the Hamilton Health Sciences Corporation andMrs. Avril Rogers for her secretarial assistance.

REFERENCES

1. Burnett AK, Grimwade D, Solomon E, Wheatley K, Goldstone AH.Presenting white blood cell count and kinetics of molecular remissionpredict prognosis in acute promyelocytic leukemia treated withall-trans-retinoic acid: result of the randomized MRC trial. Blood 1999;93:4131–4143.

2. Tallman MS. Therapy of acute promyelocytic leukemia:all-trans-retinoic acid and beyond. Leukemia 1988;12:S37–S40.

3. Lo Coco F, Nervi C, Avvisati G, Mandelli F. Acute promyelocyticleukemia: a curable disease. Leukemia 1998;12:1866–1880.

4. Tallman MS, Andersen JW, Schiffer CA, Appelbaum FR, Feusner JH,Ogden A, Shepherd L, Willman C, Bloomfield CD, Rowe JM, WiernikPH.all-trans-Retinoic acid in acute promyelocytic leukemia. N Engl JMed 1997;337:1021–1028.

5. Fenaux P, Chastang C, Chevret S, Sanz M, Dombret H, ArchimbaudE, Fey M, Rayon C, Huguet F, Sotto JJ, Gardin C, Makhoul PC,

CD34-Positive APL 39

Page 7: CD34-positive acute promyelocytic leukemia is associated with leukocytosis, microgranular/hypogranular morphology, expression of CD2 and bcr3 isoform

Travade P, Solary E, Fegueux N, Bordessoule D, Miguel JS, Link H,Desablens B, Stamatoullas A, Deconinck E, Maloisel F, Castaigne S,Preudhomme C, Degos L. A randomized comparison ofall-trans-retinoic acid (ATRA) followed by chemotherapy and ATRA plus che-motherapy and the role of maintenance therapy in newly diagnosedacute promyelocytic leukemia. Blood 1999;94:1192–1200.

6. Diverio D, Rossi V, Avvisati G, De Santis S, Pistilli A, Pane F, SaglioG, Martinelli G, Petti MC, Santoro A, Pelicci PG, Madelli F, Biondi A,Lo Coco F. Early detection of relapse by prospective reverse transcrip-tase-polymerase chain reaction analysis of the PML-RARa fusiongene in patients with acute promyelocytic leukemia enrolled in theGIMEMA-AIEOP multicentre “AIDA” trial. Blood 1998;92:784–789.

7. Asou N, Adachi K, Tamura J, Kanamaru A, Kageyama S, Hiraoka A,Omoto E, Akiyama H, Tsubaki K, Saito K, Kuriyama K, Oh H, KitanoK, Miyawaki S, Takeyama K, Yamada O, Nishikawa K, Takahashi M,Matsuda S, Ohtake S, Suzushima H, Emi N, Ohno R. Analysis ofprognostic factors in newly diagnosed acute promyelocytic leukemiatreated withall-trans-retinoic acid and chemotherapy. J Clin Oncol1998;16:78–85.

8. Chou WC, Tang JL, Yao M, Liang YJ, Lee FY, Lin MT, Wang CH,Shen MC, Chen YC, Tien HF. Clinical and biological characteristicsof acute promyelocytic leukemia in Taiwan: a high relapse rate inpatients with high initial and peak white blood cell counts duringall-trans-retinoic acid treatment. Leukemia 1997;11:921–928.

9. Slack JL, Arthur DC, Lawrence D, Mrozek K, Mayer RJ, Davey FR,Tantravahi R, Pettenati MJ, Bigner S, Carroll AJ, Rao KW, SchifferCA, Bloomfield CD. Secondary cytogenetic changes in acute promy-elocytic leukemia—prognostic importance inpatients treated with che-motherapy alone and association with the intro 3 breakpoint of thePML gene: a cancer and leukemia group B study. J Clin Oncol 1997;15:1786–1795.

10. Gallagher RE, Willman CL, Slack JL, Andersen JW, Li YP,Viswanatha D, Bloomfield CD, Appelbaum FR, Schiffer CA, TallmanMS, Wiernik PH. Association of PML-RARa fusion mRNA type withpretreatment hematologic characteristics but not treatment outcome inacute promyelocytic leukemia: an intergroup molecular study. Blood1997;90:1656–1663.

11. Fukutani H, Naoe T, Ohno R, Yoshida H, Miyawaki S, Shimazaki C,Miyake T, Nakayama Y, Kobayashi H, Goto S, Takeshita A, Kobaya-shi S, Kato Y, Shiraishi K, Sasada M, Ohtake S, Murakami H, Ko-bayashi M, Endo N, Shindo H, Matsushita K, Hasegawa S, Tsuji K,Ueda Y, Tominaga N, Furuya H, Inoue Y, Takeuchi J, Morishita H,Lida H. Isoforms of PML-retinoic acid receptor alpha fused transcriptsaffect neither clinical features of acute promyelocytic leukemia norprognosis after treatment withall-trans-retinoic acid. Leukemia 1995;9:1478–1482.

12. Guglielmi C, Martelli MP, Diverio D, Fenu S, Vegna ML, Cantu-Rajnoldi A, Biondi A, Cocito MG, Vecchio LD, Tabilio A, Avvisati G,Basso G, Lo Coco F. Immunophenotype of adult and childhood acutepromyelocytic leukaemia: correlation with morphology, type of PMLgene breakpoint and clinical outcome. A cooperative Italian study on196 cases. Br J Haematol 1998;102:1035–1041.

13. De Rossi G, Avvisati G, Coluzzi S, Fenu S, LoCoco F, Lopez M,Nanni M, Pasqualetti D, Mandelli F. Immunological definition ofacute promyelocytic leukemia (FAB M3): a study of 39 cases. Eur JHaematol 1990;45:168–171.

14. Paietta E, Andersen J, Gallagher R, Bennett J, Yunis J, Cassileth P,Rowe J, Wiernik PH. The immunophenotype of acute promyelocyticleukemia (APL): an ECOG study. Leukemia 1994;8:1108–1112.

15. Neame PB, Soamboonsrup P, Leber B, Carter RF, Sunisloe L, Patter-son W, Orzel A, Bates S, McBride JA. Morphology of acute promy-elocytic leukemia with cytogenetic or molecular evidence for the di-agnosis: characterization of additional microgranular variants. Am JHematol 1997;56:131–142.

16. Rovelli A, Biondi A, Rajnoldi AC, Conter V, Giudici G, Jankovic M,Locasciulli A, Rizzari C, Romitti L, Rossi MR, Schiro R, Tosi S,

Uderzo C, Masera G. Microgranular variant of acute promyelocyticleukemia in children. J Clin Oncol 1992;10:1413–1418.

17. Biondi A, Luciano A, Bassan R, Mininni D, Specchia G, Lanzi E,Castagna S, Cantu-Rajnoldi A, Liso V, Masera G, Barbui T, RambaldiA. CD2 expression in acute promyelocytic leukemia is associated withmicrogranular morphology (FAB M3v) but not with any PML genebreakpoint. Leukemia 1995;9:1461–1466.

18. Claxton DF, Reading CL, Nagarajan L, Tsujimoto Y, Andersson BS,Estey E, Cork A, Huh YO, Trujillo J, Deisseroth AB. Correlation ofCD2 expression with PML gene breakpoints in patients with acutepromyelocytic leukemia. Blood 1992;80:582–586.

19. Ferrara F, Morabito F, Martino B, Specchia G, Liso V, Nobile F,Boccuni P, Di Noto R, Pane F, Annunziata M, Schiavone EM, DeSimone M, Guglielmi C, Del Vecchio L, Lo Coco F. CD56 expressionis an indicator of poor clinical outcome in patients with acute promy-elocytic leukemia treated with simultaneousall-trans-retinoic acid andchemotherapy. J Clin Oncol 2000;18:1295–1300.

20. Piedras J, Lopez-Karpovitch X, Cardenas R. Light scatter and immu-nophenotypic characteristics of blast cells in typical acute promyelo-cytic leukemia and its variant. Cytometry 1998;32:286–290.

21. Casasnovas RO,Campos L, Mugneret F, Charrin C, Bene MC, GarandR, Favre M, Sartiaux C, Chaumarel I, Bernier M, Faure G, Solary E.Immunophenotypic patterns and cytogenetic anomalies in acute non-lymphoblastic leukemia subtypes: a prospective study of 432 patients.Leukemia 1998;12:34–43.

22. Edwards RH, Wasik MA, Finan J, Rodriguez R, Moore J, Kamoun M,Rennert H, Bird J, Nowell PC, Salhany KE. Evidence for early hema-topoietic progenitor cell involvement in acute promyelocytic leukemia.Am J Clin Pathol 1999;112:819–827.

23. Raspadori D, Lauria F, Ventura MA, Rondelli D, Visani G, de Vivo A,Tura S. Incidence and prognostic relevance of CD34 expression inacute myeloblastic leukemia: analysis of 141 cases. Leuk Res 1997;21:603–607.

24. Lauria F, Raspadori D, Rondelli D, Ventura MA, Fiacchini M, VisaniG, Forconi F, Tura S. High bcl-2 expression in acute myeloid leukemiacells correlates with CD34 positivity and complete remission rate.Leukemia 1997;11:2075–2078.

25. Weick JK, Kopecky KJ, Appelbaum FR, Head DR, Kingsbury LL,Balcerzak SP, Bickers JN, Hynes HE, Welborn JL, Simon SR, GreverM. A randomized investigation of high-dose versus standard-dose cy-tosine arabinoside with daunorubicin in patients with previously un-treated acute myeloid leukemia: a southwest oncology group study.Blood 1996;88:2841–2851.

26. Warrell RP. Retinoid resistance in acute promyelocytic leukemia: newmechanisms, strategies and implications. Blood 1993;82:1949–1953.

27. Miyauchi J, Inatomi Y, Ohyashiki K, Asada M, Mizutani S, ToyamaK. The in vitro effects ofall-trans-retinoic acid and hematopoieticgrowth factors on the clonal growth and self-renewal of blast stemcells in acute promyelocytic leukemia. Leuk Res 1997;21:285–294.

28. Larghero J, Rousselot P, Miclea JM, Solary E, Chomienne C. Differ-ential effects ofall-trans-retinoic acid on acute promyelocytic leuke-mia cells related to the expression of CD34 antigen. Br J Haematol1998:102:100a.

29. Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DAG, Gral-nick HR, Sultan C. Proposals for the classification of the acute leu-kaemias. Br J Haematol 1976;33:451–458.

30. Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DAG, Gral-nicks HR, Sultan C. A variant form of hypergranular promyelocyticleukemia (M3). Br J Haematol 1980;44:169–170.

31. McKenna RW, Parkin J, Bloomfield CD, Sundberg RD, Brunning RD.Acute promyelocytic leukaemia: a study of 39 cases with identificationof a hyperbasophilic microgranular variant. Br J Haematol 1982;50:201–214.

32. Gallagher RE, Li YP, Rao S, Paietta E, Andersen J, Etkind P, BennettJM, Tallman MS, Wiernik PH. Characterization of acute promyelo-cytic leukemia cases with PML-RARa break/fusion sties in PML exon

40 Foley et al.

Page 8: CD34-positive acute promyelocytic leukemia is associated with leukocytosis, microgranular/hypogranular morphology, expression of CD2 and bcr3 isoform

6: identification of a subgroup with decreased in vitro responsivenessto all-trans-retinoic acid. Blood 1995;86:1540–1547.

33. Foley R, Soamboonsrup P, Kouroukis T, Leber B, Carter RF, SunisloeL, Chorneynko K, Neame PB. PML-RARa APL with undifferentiatedmorphology and stem cell immunophenotype. Leukemia 1998;12:1492–1493.

34. Cunningham I, Gee TS, Reich LM, Kempin SJ, Naval AN, ClarksonB. Acute promyelocytic leukemia: treatment results during a decade atMemorial Hospital. Blood 1989;73:1116–1122.

35. Rodeghiero F, Avvisati G, Castaman G, Barbui T, Mandelli F. Earlydeaths and anti-hemorrhagic treatments in acute promyelocytic leuke-mia. A GIMEMA retrospective study in 268 consecutive patients.Blood 1990;75:2112–2117.

36. Tallman MS, Kwaan HC. Reassessing the Hemostatic disorder asso-ciated with acute promyelocytic leukemia. Blood 1992;79:543–553.

37. Barbui T, Finazzi G, Falanga A. The impact ofall-trans-retinoic acidon the coagulopathy of acute promyelocytic leukemia. Blood 1998;91:3093–3102.

38. Biondi A, Rovelli A, Cantu-Rajnoldi C, Fenu S, Basso G, Luciano A,Rondelli R, Mandelli F, Masera G, Testi AM. Acute promyelocyticleukemia in children: experience of the Italian Pediatric Hematologyand Oncology Group (AIEOP). Leukemia 1994;8:1264–1268.

39. Albitar M, Chang KS, Pierce S, Kantarjian H, Estey E. The short formof PML-RARa fusion transcript is associated with poor survival. LeukRes 1999;23:89–90.

40. Huang W, Sun GL, Li XS, Cao Q, Lu Y, Jang GS, Zhang FQ, Chai JR,Wang ZY, Waxman S, Chen Z, Chen SJ. Acute promyelocytic leuke-mia: clinical relevance of two major PML-RARa isoforms and detec-tion of minimal residual disease by retrotranscriptase/polymerasechain reaction to predict relapse. Blood 1993;82:1264–1269.

41. Vahdat L, Masiak P, Miller WH, Eardley A, Heller G, Scheinberg DA,Warrell RP. Early mortality and the retinoic acid syndrome in acutepromyelocytic leukemia: impact of leukocytosis, low-dose chemo-therapy, PMN/RARa isoform, and CD13 expression in patients treatedwith all-trans-retinoic acid. Blood 1994;84:3843–3849.

42. Muindi J, Franel SR, Miller WH, Jakubowski A, Scheinberg DA,Young CW, Dmitrovsky E, Warrell RP. Continuous treatment with

all-trans-retinoic acid causes a progressive reduction in plasma drugconcentrations: implications for relapse and retinoid “resistance” inpatients with acute promyelocytic leukemia. Blood 1992;79:299–303.

43. Delva L, Cornic M, Balitrand N, Guidez F, Miclea JM, Delmer A,Teillet F, Fenaux P, Castaigne S, Degos L, Chomienne C. Resistanceto all-trans-retinoic acid (ATRA) therapy in relapsing acute promy-elocytic leukemia: study of in vitro ATRA sensitivity and cellularretinoic acid binding protein levels in leukemic cells. Blood 1993;82:2175–2181.

44. Kizaki M, Ueno H, Yamazoe Y, Shimada M, Takayama N, Muto A,Matsushita H, Nakajima H, Morikawa M, Koeffler HP, Ikeda Y.Mechanisms of retinoid resistance in leukemia cells: possible role ofcytochrome P450 and P-glycoprotein. Blood 1996;87:725–733.

45. Shao W, Benedetti L, Lamph WW, Nervi C, Miller WH. A retinoid-resistant acute promyelocytic leukemia subclone expresses dominantnegative PML-RARa mutation. Blood 1997;89:4282–4289.

46. Imaizumi M, Suzuki H, Yoshinari M, Sato A, Saito T, Sugawara A,Tsuchiya S, Hatae Y, Fujimoto T, Kakizuka A, Konno T, Iinuma K.Mutations in the E-domain of RARa portion of the PML-RARa chi-meric gene may confer clinical resistance toall-trans-retinoic acid inacute promyelocytic leukemia. Blood 1998;92:374–382.

47. Nason-Burchenal K, Allopenna J, Begue A, Stehelin D, Dmitrovsky E,Martin P. Targeting of PML-RARa is lethal to retinoic acid-resistantpromyelocytic leukemia cells. Blood 1998;92:1758–1767.

48. Ding W, Li YP, Nabile LM, Grills G, Carrera I, Paietta E, TallmanMS, Wiernik PH, Gallagher RE. Leukemic cellular retinoic acid re-sistance and missense mutation in the PML-RARa fusion gene afterrelapse of acute promyelocytic leukemia from treatment withall-trans-retinoic acid and intensive chemotherapy. Blood 1998;92:1172–1183.

49. Purton LE, Bernstein ID, Collins SJ.all-trans-Retinoic acid delays thedifferentiation of primitive hematopoietic precursors (lin- c-kit + sca-1+) while enhancing the terminal maturation of committed granulo-cyte/monocyte precursors. Blood 1999;94:483–495.

50. Bradbury DA, Aldington S, Zhu Y-M, Russell NH. Down-regulationof bcl-2 in AML blasts byall-trans-retinoic acid and its relationship toCD34 antigen expression. Br J Haematol 1996;94:671–675.

CD34-Positive APL 41