A distinctive nuclear morphology in acute myeloid leukemia is ... - Nature

Leukemia (2004) 18, 1591–1598 & 2004 Nature Publishing Group All rights reserved 0887-6924/04 $30.00 www.nature.com/leu

A distinctive nuclear morphology in acute myeloid leukemia is strongly associated with loss of HLA-DR expression and FLT3 internal tandem duplication SJ Kussick1, DL Stirewalt2, HS Yi1, KM Sheets2, E Pogosova-Agadjanyan2, S Braswell3, TH Norwood3, JP Radich2 and BL Wood1 1 Department of Laboratory Medicine, University of Washington, Seattle, WA, USA; 2The Fred Hutchinson Cancer Research Center, Seattle, WA, USA; and 3Department of Pathology, University of Washington, Seattle, WA, USA

In a 5-year survey of nonpromyelocytic/nonmonocytic acute myeloid leukemias (AMLs) diagnosed in the University of Washington Hematopathology Laboratory, we identified 19 cases containing distinctive, cup-like nuclear indentation in 10% or more of the blasts (‘AML-cuplike’). Fourteen of these cases (74%) demonstrated near-complete loss of HLA-DR expression, while the other five cases showed partial loss of HLA-DR. A total of 16 of the cases (84%) demonstrated internal tandem duplication (ITD) of the Flt3 gene. When compared to a selected set of AMLs lacking this nuclear morphology, AMLcuplike was significantly more likely to lack HLA-DR and CD34 expression, to express CD123 without CD133, to have a normal karyotype, and to harbor the Flt3 ITD. To characterize AMLcuplike in an unselected series of AMLs, we analyzed 42 consecutive nonpromyelocytic/nonmonocytic AMLs diagnosed in our laboratory during a 6-month period in 2002. Strikingly, in this unselected series, there was a statistically significant coincidence of invaginated nuclear morphology, loss of HLADR, and presence of the Flt3 ITD beyond that expected if these three features were unrelated, suggesting that AMLs with these three features may represent a distinct AML subset. Leukemia (2004) 18, 1591–1598. doi:10.1038/sj.leu.2403458 Published online 2 September 2004 Keywords: acute myeloid leukemia; Flt3; CD123; CD133; HLA-DR; nuclear invagination

Introduction Current classification of acute myeloid leukemia (AML) under the World Health Organization (WHO) system relies on combining information about the morphologic, immunophenotypic, and molecular/cytogenetic features of the leukemic cells.1 The combination of all three types of information is typically used to diagnose acute promyelocytic leukemia (APL), while the combination of morphology and immunophenotype readily identifies monocytic leukemias. Although some unique forms of the nonpromyelocytic/nonmonocytic acute myeloid leukemias have been identified, such as AML associated with the t(8;21)(q22;q22), there are likely to be a number of biologically distinct subsets of these AMLs remaining to be discovered. The most common molecular defect identified to date in AML is internal tandem duplication (ITD) in the juxtamembraneencoding region of the Flt3 gene, which is present in 20–30% of AMLs,2–4 and leads to ligand-independent, constitutive activation of the Flt3 tyrosine kinase function.5,6 The majority of clinical studies of the Flt3 ITD in both adults7–12 and children13,14 have shown that the presence of a single Flt3 ITD allele is an independent poor prognostic factor in AML, particularly in nonpromyelocytic AML. However, one of these Correspondence: Dr S Kussick, Laboratory Medicine, University of Washington Medical Center, Box 357110, 1959 NE Pacific Street, Seattle, WA 98195, USA; Fax: þ 206 598 6189; E-mail: [email protected] Received 1 December 2003; accepted 4 June 2004; Published online 2 September 2004

studies8 suggested that the presence of the Flt3 ITD only tends to influence prognosis when the PCR-defined ratio of the ITD allele to the wild-type allele (the Flt3 allelic ratio) is relatively high, while another study found that the adverse prognostic impact of the ITD is only observed if both wild-type alleles are affected.15 Finally, some studies have suggested that Flt3 ITDs may not be of prognostic significance in all types of AML.16 Together, these clinical findings suggest that Flt3 ITD-positive AMLs represent a heterogeneous population, and may contain subgroups with relatively favorable and unfavorable prognoses. We now describe a distinctive group of 19 AML cases in which the blasts exhibited prominent nuclear invagination with frequent cup-like nuclear indentation, relatively scanty cytoplasm with little granularity, and no evidence of the t(15;17) despite low-to-absent expression of HLA-DR and CD34 in most cases. While there have been very limited prior descriptions of similar AML cases, we have gone beyond these prior descriptions to demonstrate uniform expression of CD123, low-tonegative CD133, and the presence of the Flt3 ITD among a very high percentage of these AMLs. When an unselected series of consecutive AMLs diagnosed in our laboratory during a 6-month period was reviewed, there was a statistically significant coincidence of invaginated nuclear morphology, loss of HLADR, and the Flt3 ITD. Our findings raise the possibility that these cases represent a unique subset of AML, although clinical follow-up of a larger series of cases will be required to confirm this.

Materials and methods

Case identification We first identified sporadic cases of AML in which a number of the blasts had a distinctive, cup-like nuclear invagination. We then performed a retrospective examination of the University of Washington Hematopathology Laboratory (UWHL) database to identify similar cases. Since nuclear invagination is a common feature of microgranular APL as well as the myelomonocytic and monocytic AMLs, we limited the cases selected to the nonpromyelocytic and nonmonocytic AMLs. In addition to being nonpromyelocytic and nonmonocytic, cases chosen for the series demonstrated 10% of more blasts with prominent, cup-like nuclear invagination spanning at least 25% of the nuclear diameter, based on the independent morphologic evaluation of two hematopathologists (SJK and BLW). Cases satisfying these criteria, which constitute Series 1 of this study, are described in Table 1. Note that loss of HLA-DR was not a criterion for choosing cases for this series, although in all of these cases, 20% or more of the myeloid blasts demonstrated loss of surface expression of the myeloid blast-associated antigen HLA-DR by flow cytometry, compared to a control antibody, usually an isotype-matched control. Cases included in

FLT3 internal tandem duplication and AML morphology SJ Kussick et al

1592 Series 1 included de novo AML, relapsed AML, and AML with underlying morphologic evidence of myelodysplasia. After Series 1 was established, two additional series of nonpromyelocytic/nonmonocytic AMLs diagnosed in our laboratory (Series 2 and 3) were identified to help evaluate the features in Series 1. Series 2 included 24 noninvaginated AMLs diagnosed during the same 5-year period as Series 1, and subject to the same extensive flow cytometric and molecular analyses as the cases in Series 1. Series 3 included all AML cases diagnosed in our laboratory between March 2002 and September 2002, and was created to enable the assessment of the three key features identified in Series 1 – invaginated nuclear morphology, loss of surface HLA-DR, and presence of the Flt3 ITD – in a group of unselected AMLs. For the few patients having multiple specimens involved by AML during this 6-month period, the specimen with the highest blast percentage was included in Series 3. This study was approved by the Human Subjects Review Committee of the University of Washington (Application #02-2717-E 01).

t(15;17) were as follows: external RARr-ex (50 0 0 CAGTTCTTGTCCCGGTGACAC-3 ) and PMLf-ex (5 -CAGCGCGACTACGAGGAGAT-30 ); internal RARr-in (50 -ACAGGCGCTGACCCCATAGT-30 ) and PMLf-in (50 -CGATGGCTTCGACGAGT TCA-30 ).

Statistics Statistical comparisons between the novel AMLs and other control AMLs were performed as follows: (1) continuous numerical data, such as mean patient ages among the case and comparison populations, were compared using the twosided Student t-test; (2) proportions, such as the percentage of cytogenetically abnormal cases among the Series 1 and Series 2 populations, were compared using the w2 test based on the appropriate number of degrees of freedom. For all statistical tests, statistical significance was defined by a P-value of 0.05 or less.

Results

Flow cytometry Four-color flow cytometry was performed on Coulter XL instruments (Hialeah, FL, USA), using commercially available reagents as previously described.17 In several cases, the evaluations of CD123 (clone 9F5, Becton-Dickinson (BD), San Jose, CA, USA), CD133 (clone AC133, Miltenyi Biotec, Bergisch Gladbach, Germany), and CD135 (clone 4D8, BD) were performed on cryopreserved material; in such analyses, nonviable cells were excluded from the analysis by virtue of their decreased forward light scatter. The flow cytometry data were analyzed using software developed in the UWHL (BLW). Antigen expression was regarded as ‘negative’ on the blasts when 10% or fewer of the blasts showed fluorescence intensities in excess of the fluorescence intensity associated with an isotype-matched internal negative control antibody.

Molecular and cytogenetic methods DNA-based PCR was used to screen for Flt3 ITDs. Most of the PCR amplifications used a hex-labeled forward primer and an unlabeled reverse primer based on published sequences;11 in a subset of the PCR reactions, a ned-labeled reverse primer was used along with the hex-labeled forward primer. The PCR products were separated by capillary electrophoresis on the Applied Biosystems ABI 310 instrument (Foster City, CA, USA), and GeneScan software was used to evaluate the size of the product(s). Single-stranded conformational polymorphism (SSCP) analysis followed by DNA sequencing were used to confirm most Flt3 ITDs, and to rule out point mutations in exons 14, 15, and 20, as previously described.16 Novel primers for Flt3 exons 14 and 15 were developed for this purpose, as previously described.18 Cytogenetic studies were performed on G-banded preparations.19 Fluorescent in situ hybridization (FISH) to detect the t(15;17) was performed using probes for the promyelocytic leukemia (PML) gene and the retinoic acid receptor-alpha (RARA) gene at 15q22 and 17q21, respectively (Vysis, Downers Grove, IL, USA), according to the manufacturer’s instructions. Combined reverse transcription-PCR (RT-PCR) for the t(15;17) and t(17;15) were performed using a previously described nested technique starting with 1 mg of total RNA.20 Primers for the t(17;15) assay have previously been described,21 and for the Leukemia

Patient characteristics A total of 19 cases of AML in which 10% of more of the blasts had prominent, cup-like nuclear invagination spanning at least 25% of the nuclear diameter were identified independently by two hematopathologists (SJK and BLW). For brevity, these cases are heretofore referred to as ‘AML-cuplike’, and the overall group of 19 cases as Series 1. Since nuclear invagination is a common feature of microgranular APL as well as the myelomonocytic and monocytic AMLs, we limited the cases selected to the nonpromyelocytic and nonmonocytic AMLs. The patients in Series 1 represented about 3.8% of all AML patients diagnosed in the University of Washington Hematopathology Laboratory between mid-1998 and mid-2003. The characteristics of these patients are summarized in Table 1. Table 2A compares a variety of features of AML-cuplike to those in a selected series of 24 cases of AML (Series 2) diagnosed during the same time period and subjected to the same detailed immunophenotypic and molecular analyses as the AML-cuplike cases, but lacking this distinctive nuclear morphology. Overall, the AML-cuplike patients were older than the patients in Series 2. In Table 2B, several of the older patients in Series 1, and several of the younger patients in Series 2, were excluded from the analysis, such that the mean ages of the two comparison populations were virtually identical, effectively making this an age-adjusted comparison. In both Table 2A and B, AML-cuplike cases included a higher percentage of female subjects, but this tendency did not reach statistical significance.

Morphology Representative examples of the nuclear morphology in AMLcuplike cases are demonstrated in Figure 1. In all cases, at least 10% of the blasts had a prominent, cup-like nuclear invagination spanning at least 25% of the nuclear diameter (Cases 4 and 11). In addition, in some of the cases, a number of the blasts also had a bilobed nuclear appearance in which a cleft could be seen spanning the entire diameter of the nucleus (Case 7). However, in the latter cases, there were also at least 10% blasts containing cup-like morphology. AMLs in which the blasts showed irregular nuclear contours, or even prominent nuclear clefts, but without cup-like morphology were not included in Series 1.


1593 Table 1

Summary of AML-cuplike cases (Series 1)

Case

Age

Sex

Source

% HLA-DR+ among blastsa

% CD34+ among blastsa

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

85 74 73 88 37 68 57 79 42 81 81 49 49 80 74 68 44 68 66

F F F F M M F F F M F M F F M F F F M

PB PB PB PB BM PB PB PB PB/BM PB PB PB BM PB BM BM PB PB PB

o10 o10 o10 o10 30 o10 o10 o10 o10 o10 o10 o10 o10 o10 B60 B75 B75 B75 B75

o10 B100 o10 o10 o10 o10 B20 o10 o10 o10 o10 B50 o10 o10 B50 B50 o10 B15 o10

CD123+, CD133 blasts

FLT3 ITD

ITD:WT ratiob

Cyto genetics

t(15;17) RT-PCR or FISH

NDc ND Yes ND Yes Yes Yes Yes Yes Yes Yes Yes ND Yes Yes Yes Yes Yes Yes

Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No No Yes Yes Yes Yes No

0.09 0.33 0.86 0.66 1.16 0.73 0.8 0.42 0.44 0.28 0.94 0.77 NA NA 7.3 4 0.75 0.12 NA

Normal Normal ND ND Normal Normal ND ND Normal ND Normal Normal Normal ND Normal Normal Normal ND Inv(9)

ND ND Neg. ND ND Neg. Neg. Neg. Neg. Neg. Neg. Neg. ND Neg. Neg. ND Neg. Neg. Neg.

a

Percentages of HLA-DR-positive and CD34-positive cells among the total myeloid blast population, as determined by four-color flow cytometry. Cases in which less than 10% of the blasts expressed HLA-DR or CD34 were considered negative for these antigens. b Denotes ratio of Flt3 ITD allele-associated PCR peak fluorescence to wild-type Flt3 allele-associated PCR peak fluorescence (Flt3 allelic ratio), normalized to the percentage of myeloid blasts in the specimen. c ND, not done.

Table 2 (A) Comparison of full AML-cuplike series with a selected series of other nonpromyelocytic/nonmonocytic AMLs (Series 2); (B) Ageadjusted comparison of AML-cuplike cases with Series 2 AML type

n

Mean age

Male: female

HLA-DR lossa

CD34 lossb

Minimal CD13 or CD117c

CD123+, CD133

Abnormal cytogeneticsd

FLT3-ITD+

(A) Cup-like Others

19 24

66.5* 56.0*

6:13 14:10

68%** 17%**

68%*** 8.3%***

58%**** 21%****

100%*** 39%***

8.3% (12)*** 81% (16)***

84%*** 25%***

(B) Cup-like Others

15 19

61.9 61.4

5:10 12:7

60%* 21%*

53%* 11%*

53% 21%

100%** 44%**

10% (10)** 75% (12)**

80%** 26%**

a

Percent of cases in which 10% or fewer of the leukemic blasts expressed HLA-DR. Percent of cases in which 10% or fewer of the leukemic blasts expressed CD34. c Percent of cases showing very low-to-absent expression of CD13 or CD117. d Percent of cases with abnormal conventional cytogenetics, among cases with available cytogenetic information (the latter numbers are listed in parentheses). *Statistically significant difference, Po0.05, Student’s t test. **Statistically significant difference, Po0.01, w2 test, 1 degree of freedom. ***Statistically significant difference, Po0.001, w2 test, 1 degree of freedom. ****Statistically significant difference, Po0.05, w2 test, 1 degree of freedom. b

The AML-cuplike blasts generally demonstrated relatively scanty, pale basophilic cytoplasm with little cytoplasmic granularity, and only occasional Auer rods. In most cases, the blasts showed near-uniform myeloperoxidase positivity (Figure 1, MPO), but generally did not show the intense, needle-like positivity characteristic of APL. Nonspecific esterase stains were performed on a few cases of AML-cuplike, and were negative (not shown).

Immunophenotype By flow cytometry based on CD45 vs side scatter gating, all 19 AML-cuplike cases demonstrated a predominant myeloid blast

population representing greater than 60% of the leukocytes. In 16 of the 19 cases (84%), more mature granulocytes represented 7% or fewer of the leukocytes, and blasts represented over 90% of the combined blast and granulocyte populations. Compared to the patterns of antigen expression on normal myeloid blasts (Figures 2a and 3a), AML-cuplike blasts demonstrated a distinctly abnormal, immunophenotype. In 13 of the 19 cases (68%), the blasts demonstrated virtually complete loss of HLA-DR, as defined by the presence of 90% or greater HLA-DR-negative blasts, while in a 14th case (case 5), 70% of the blasts were HLA-DR negative. The five remaining AML-cuplike cases demonstrated loss of HLA-DR expression on 24–40% of the myeloid blasts (partial loss of HLA-DR). Leukemia


1594 demonstrated strong expression of CD123 (IL-3a receptor), little or no CD133 (an early myeloid blast associated-antigen), and a low-level of Flt3, or CD135, on the AML-cuplike blasts (Figures 3b and c). As shown in Table 2A and B, AML-cuplike was significantly more likely to express CD123 without CD133 than the Series 2 cases, independent of age adjustment. To summarize the flow cytometric data, AML-cuplike shows essentially uniform expression of CD123, with frequent loss of HLA-DR and CD34 expression, and a virtual absence of CD133. The CD34-negative cases of AML-cuplike are more likely to show abnormally low-level expression of CD13 and CD117, and aberrant expression of CD56, compared to the CD34positive cases of AML-cuplike.

Cytogenetic and molecular features

Figure 1 AML-cuplike demonstrates prominent nuclear invagination. Wright–Giemsa peripheral blood smears from cases 4 and 11 demonstrate the cup-like nature of the nuclear invaginations in many of the blasts, whereas the image of case 7 demonstrates the frank nuclear clefts seen in some blasts (arrowheads), which presumably represents an extreme degree of nuclear invagination. The blasts in AML-cuplike are myeloperoxidase (MPO) positive (lower right panel, case 11), but usually do not show the intense, needle-like positivity, characteristic of acute promyelocytic leukemia.

Of the AML-cuplike cases, 13 (68%) lacked CD34 (present on less than 10% of the blasts compared to an isotype-matched control antibody). Interestingly, the CD34-negative (Figures 2b and 3b) and CD34-positive cases (Figures 2c and 3c) had somewhat distinctive immunophenotypes. The pan-myeloid antigen CD13 and the myeloid blast-associated antigen CD117 (c-kit) were each expressed at very low-to-absent levels in the majority of CD34-negative cases, while none of the CD34-positive cases showed very low-to-absent levels of CD13 or CD117. In addition, six of the 13 CD34-negative cases showed aberrant coexpression of CD56 on at least a subset of the blasts (Figure 2b), while only one of the six CD34-positive cases showed CD56 expression on the blasts (this case not shown). In comparison with the 24 cases of AML in the full Series 2 (Table 2A), the AML-cuplike cases were significantly more likely to be HLA-DR negative and CD34 negative, and also more likely to express very low-to-absent CD13 or CD117 (see Table 2A). When the immunophenotypes of the ageadjusted groups were compared (Table 2B), AML-cuplike cases were still significantly more likely to lack HLA-DR and CD34. In all, 14 cases of AML-cuplike, including eight CD34negative and six CD34-positive cases, were subjected to additional flow cytometric immunophenotyping. This analysis Leukemia

AML-cuplike cases were significantly more likely to have a normal karyotype by conventional cytogenetics than the AMLs in Series 2, even after age adjustment (Tables 2A and B). Although lack of HLA-DR and CD34 expression is particularly associated with APL, and the invaginated nuclear morphology of AML-cuplike somewhat resembles the nuclear bilobation seen in many cases of microgranular APL, there was no evidence of the t(15;17) in the 12 AML-cuplike cases analyzed by conventional cytogenetics (Table 1). Moreover, none of the cases analyzed by FISH (six cases) or RT-PCR (seven cases), which included six of the seven cases without cytogenetics, demonstrated the t(15;17). Only one of the 19 AML-cuplike cases, involving the oldest patient in the series (case 4, 88 years old), did not have cytogenetic or RT-PCR/FISH analysis, but was retained in the series because of its distinctive nuclear morphology and immunophenotype (see Figure 1). Since AML-cuplike demonstrates some features associated with the Flt3 ITD, including a high frequency of normal cytogenetics, PCR analysis of the Flt3 gene was performed in all of these cases, and in all comparison cases. Strikingly, 16 of the 19 AML-cuplike cases (84%) demonstrated an ITD, vs only 25% of the AMLs in Series 2 (Table 2A), and this significant difference persisted in the age-adjusted comparison (Table 2B). In all ITD-positive cases, the presence of the Flt3 ITD was confirmed by either SSCP and DNA sequencing (majority of cases) or the presence of the same sized ITD using a ned-labeled reverse primer (in addition to the hex-labeled forward primer). Among the AML cup-like cases, the nonadjusted ratios of the Genescan fluorescence associated with the Flt3 ITD-containing allele to the fluorescence associated with the wild-type allele (the Flt3 allelic ratio) ranged from 0.08 in case 1 to 5.6 in case 4, with a mean of 0.99 and a median of 0.65. After normalization to the percentage of blasts in each case, the Flt3 allelic ratios varied from 0.09–7.0, with a mean of 1.23 and a median of 0.74. The six Flt3 ITD-positive cases in the Series 2 comparison group had nonadjusted Flt3 allelic ratios ranging from 0.13 to 5.6 (mean of 1.8 and median of 0.88), and normalized ratios ranging from 0.22 to 6.0 (mean of 2.6 and median of 2.1). However, none of these differences in the mean allelic ratios reach statistical significance, nor did the differences in the allelic ratios in the age-matched comparison. The only feature of the AML-cuplike cases that showed some relationship to the Flt3 allelic ratio was the degree of HLA-DR expression on the blasts. Specifically, the three AML-cuplike cases with the highest Flt3 allelic ratios all had some level of HLA-DR expression on the blasts (30% or more HLA-DRpositive blasts), such that the mean normalized allelic ratio was 2.67 among the five Flt3 ITD-positive cases with some HLA-DR


1595

Figure 2 AML-cuplike has a distinctive immunophenotype, part 1. The images show two-dimensional histograms (‘dot-plots’) with the results of four-color flow cytometric evaluation of HLA-DR, CD33, CD13, CD19, CD56, CD38, and CD34. Normal bone marrow is shown in the top row (a), a case with uniformly CD34-negative blasts in the middle row (b, case 6), and a case with CD34 expression on a subset of the blasts in the bottom row (c, case 12). The myeloid blasts were identified by CD45 vs side scatter gating (not shown). Fluorescence intensities for all of the bound antibodies are presented on a logarithmic scale. Fluorochromes are indicated as follows: FITC, fluorescein isothiocyanate; PE, phycoerythrin; ECD, PE-Texas Red; PC5, PE-Cy5. In the CD34-negative case in part(b), note the decreased level of CD13 and the aberrant expression of CD56, which were both more characteristic of these cases than of those with CD34-positive blasts.

Figure 3 AML-cuplike has a distinctive immunophenotype, part 2. The images show two-dimensional histograms (‘dot-plots’) with the results of four-color flow cytometric evaluation of CD38, CD117, CD123, CD133, and CD135. In the CD34-negative case in part b, note the decreased level of CD117 expression, which was more characteristic of these cases than of those with CD34-positive blasts (part c). The levels of expression of CD123, CD133, and CD135 are characteristic of all cases in which these antigens were evaluated by flow cytometry. Leukemia


1596 Table 3

Observed Expected

Evaluation of all 42 nonpromyelocytic/nonmonocytic AMLs in an unselected series diagnosed from March 2002 to September 2002 Invag. nuclei

HLA-DRloss

FLT3ITD+

Invag’d nuclei + HLA-DR-loss

Invag’d nuclei + FLT3-ITD+

HLA-DR-loss+ FLT3-ITD+

Invag’d nuclei+ HLA-DR-loss+ FLT3-ITD+

11.9% (5) F

14.3% (6) F

28.6% (12) F

7.1%* (3) 1.7%

11.9%** (5) 3.4%

7.1% (3) 4.1%

7.1%*** (3) 0.49%

*Statistically significant difference, Po0.025, Chi-square test, 1 degree of freedom. **Statistically significant difference, Po0.005, Chi-square test, 1 degree of freedom. ***Statistically significant difference, Po0.025, Chi-square tests, 3 degrees of freedom.

positivity, vs a mean normalized allelic ratio of only 0.57 among the 11 Flt3 ITD-positive cases completely lacking HLA-DR. Otherwise, there was otherwise no obvious association between the Flt3 allelic ratio and specific features of AML-cuplike. Among all AML-cuplike cases tested, including the three Flt3 ITD-negative cases, no potentially activating point mutations were found in FLT3 exons 14, 15, or 20.

Evaluation of AML-cuplike in an unselected series of AMLs To characterize the relationships between invaginated nuclear morphology, loss of HLA-DR expression, and the Flt3 ITD in an unselected series of AMLs, we analyzed 42 consecutive nonpromyelocytic/nonmonocytic AMLs diagnosed in our laboratory during a 6-month period in 2002 (Series 3). As demonstrated in Table 3, based on the individual frequencies of invaginated nuclear morphology, loss of HLA-DR expression on 70% or more of the blasts, and the Flt3 ITD in this series, one can determine the expected numbers of cases showing two or three of these features in the absence of any association between these features. Among the three pairwise comparisons of these characteristics, the associations between invaginated nuclear morphology and the Flt3 ITD, and between invaginated nuclear morphology and loss of HLA-DR expression (on 70% or more of the blasts), were statistically significant. In contrast, there was no statistical association between the presence of the Flt3 ITD and loss of HLA-DR expression. Importantly, the coincidence of all three characteristics was statistically significant.

Clinical follow-up The concordance of cup-like nuclear morphology, loss of HLADR expression, and the presence of the Flt3 ITD suggested the possibility that AML-cuplike may represent a novel AML subset. However, clinical follow-up information to help determine whether these cases behaved in a biologically distinct manner was only available on six of these patients (cases 5, 6, 7, 9, 11, and 12), as the other cases were referred for diagnosis from outside institutions. Patients 5, 6, 9, and 12 were alive and in apparent remission at 17 months, 12 months, 13 months, and 4 months, respectively, following diagnosis, and patient 5 was in remission following allogeneic stem cell transplantation. Patient 7 died of pulmonary failure within 2 days of the diagnosis of AML, secondary to leukemic infiltration of the lungs (based on autopsy findings); this patient had not received induction chemotherapy. Patient 11 relapsed at 12 months from diagnosis. Leukemia

Discussion Since the Series 3 data are based on unselected AMLs, the association of cup-like nuclear invagination, loss of HLA-DR, and the presence of the Flt3 ITD does not appear to be due to selection bias, but rather appears to be a true association. Moreover, the data in Series 3 suggests that the coincidence of these three features is due to the strong associations between invaginated nuclear morphology and loss of HLA-DR, and between invaginated nuclear morphology and the Flt3 ITD, and not to a direct association of HLA-DR loss and the Flt3 ITD. This statistically significant coincidence of cup-like nuclear invagination, loss of HLA-DR expression, and presence of the Flt3 ITD in a small subset of AMLs has not been previously described. Indeed, there is only one previous suggestion of an association between this distinctive nuclear morphology and loss of HLA-DR expression in nonpromyelocytic AMLs, and this was only presented in abstract form by the CALGB Leukemia Study Group (Anastasi J, et al, Mod Pathol 2000; 13: 142A). That abstract described 74 cases of AML of the French–American– British (FAB) M1 subtype, in which 17 cases (23%) demonstrated cup-like nuclear invagination, myeloperoxidase staining, loss of HLA-DR and CD34, frequent aberrant CD56, and a normal karyotype, similar to our CD34-negative cases. The two published papers that appear to describe a small number of similar cases include a relatively recent description of HLA-DRnegative AMLs,22 which noted that three of 23 t(15;17)-negative, HLA-DR-negative cases showed nuclear folding or convolution, occurred among female subjects, and had normal karyotypes, reminiscent of AML-cuplike. Unfortunately, that study did not describe the morphology of these three cases in adequate detail to determine whether cup-like nuclear invagination was a feature. In addition, neither that study, nor the aforementioned abstract, described the expression of CD123, CD133, or CD135 in the potential AML-cuplike cases, nor characterized the mutational status of Flt3, in contrast to our study. A second paper23 that characterized Flt3 ITD-positive AMLs by flow cytometry, did note a subset of 4 HLA-DR-negative and CD34negative, non-APL cases similar to AML-cuplike, but did not describe the morphologic features of these 4 cases. In terms of the association between cup-like nuclear morphology and the Flt3 ITD, our report appears to be the first that formally links these two features. However, it should be noted that microgranular APL has frequent nuclear invagination imparting a bilobed appearance to the nucleus (but not cup-like nuclear indentation), is HLA-DR-negative, and also has a high rate of the Flt3 ITD.24–26 Therefore, among the well-characterized types of AML, microgranular APL appears to be most similar to AML-cuplike. Although the AML-cuplike and microgranular APL data suggest a general relationship between the Flt3 ITD and prominent nuclear invagination in AML, additional


1597 studies will be necessary to determine whether there is a mechanistic relationship between constitutive activation of the Flt3 tyrosine kinase and nuclear invagination, and whether the rare AML-cuplike cases without Flt3 mutations (three of 19 in our series) harbor activating mutations in other components of the Flt3 signaling pathway. While the very high frequency of the Flt3 ITD among AMLcuplike cases is most similar to the 90% frequency reported in AML bearing the t(6;9)(p23;q34),8 AML-cuplike is clearly different from AML containing the t(6;9). None of the AMLcuplike cases subjected to cytogenetic evaluation contained the t(6;9). In addition, the two cases of AML with the t(6;9) in our laboratory’s database, one of which was included as a control case in Series 2, demonstrated noninvaginated nuclear morphology and expression of HLA-DR on the majority of the blasts. Additional clinical follow-up will clearly be required to determine the prognostic significance associated with AMLcuplike, but it is intriguing that four of the six patients in our series on whom we have clinical follow-up information are alive and in apparent remission at 4–17 months after diagnosis. In their description of three HLA-DR-negative, non-APL cases with invaginated nuclear morphology reminiscent of AML-cuplike, Wetzler et al17 reported that these patients were in apparent remission at 23, 32, and 74 months after diagnosis. Therefore, we cannot exclude the possibility that AML-cuplike may have a better prognosis than other Flt3 ITD-positive AMLs, but longer follow-up of a larger series of patients will be required to examine this possibility. In summary, our findings, particularly the data in Table 3, raise the possibility that nonpromyelocytic/nonmonocytic AMLs with cup-like nuclear invagination, loss of HLA-DR expression, and the Flt3 ITD represent a rare but distinct AML subset. However, because the clinical behavior of these AMLs is not known in detail, it would be premature to conclude that these AMLs represent a distinct subtype deserving recognition under the WHO classification system. For the present, AML-cuplike would therefore be classified as ‘acute myeloid leukemia, not otherwise specified’ or ‘acute myeloid leukemia with multilineage dysplasia’ under the WHO classification, depending on the clinical and morphologic context, or as AML M1 or M2 under the FAB system. If these cases do turn out to represent a unique AML subtype, then additional features would appear to be uniform CD123 expression, low-to-negative CD34 and CD133, a tendency to have a normal karyotype, and a tendency to occur in female subjects. Our detailed description of these features represents a necessary first step toward defining the clinical behavior and prognosis of AML-cuplike, including the potential therapeutic role for Flt3 inhibitors.

Acknowledgements We thank the technical staff of the University of Washington Hematopathology Laboratory for outstanding help in the flow cytometric evaluation of all the cases in this series, and for the cryopreservation and maintenance of a number of the cases in this series. We thank the technical staff of the University of Washington Molecular Hematology Laboratory for assistance in developing the FLT3 ITD assay on the ABI 310. We thank Dr Tony Rossini of the University of Washington Department of Medical Education and Biomedical Informatics for statistical advice.This study was supported by grants K23 CA92405 and CA18029 from the National Institutes of Health and National Cancer Institute.

References 1 Jaffe ES, Harris NL, Stein H, Vardiman JW. WHO Classification of Tumours: Pathology and Genetics of Tumours of the Haematopoietic and Lymphoid Tissues. IARC Press: Lyon, 2001. 2 Gilliland DG, Griffin JD. The roles of FLT3 in hematopoiesis and leukemia. Blood 2002; 100: 1532–1542. 3 Stirewalt DL, Radich JP. The role of FLT3 in haematopoietic malignancies. NatRev Cancer 2003; 3: 650–665. 4 Nakao M, Yokota S, Iwai T, Kaneko H, Horiike S, Kashima K et al. Internal tandem duplication of the Flt3 gene found in acute myeloid leukemia. Leukemia 1996; 10: 1911–1918. 5 Kiyoi H, Towatari M, Yokota S, Hamaguchi M, Ohno R, Saito H et al. Internal tandem duplication of the FLT3 gene is a novel modality of elongation mutation which causes constitutive activation of the product. Leukemia 1998; 12: 1333–1337. 6 Hayakawa F, Towatari M, Kiyoi H, Tanimoto M, Kitamura T, Saito H et al. Tandem-duplicated Flt3 constitutively activates STAT5 and MAP kinase and introduces autonomous cell growth in IL-3dependent cell lines. Oncogene 2000; 19: 624–631. 7 Schnittger S, Schoch C, Dugas M, Kern W, Staib P, Wuchter C et al. Analysis of FLT3 length mutations in 1003 patients with acute myeloid leukemia: correlation to cytogenetics, FAB subtype, and prognosis in the AMLCG study and usefulness as a marker for the detection of minimal residual disease. Blood 2002; 100: 59–66. 8 Thiede C, Steudel C, Mohr B, Schaich M, Schakel U, Platzbecker U et al. Analysis of FLT3-activating mutations in 979 patients with acute myelogenous leukemia: association with FAB subtypes and identification of subgroups with poor prognosis. Blood 2002; 99: 4326–4335. 9 Kiyoi H, Naoe T, Nakano Y, Yokota S, Minami S, Miyawaki S et al. Prognostic implication of FLT3 and N-RAS gene mutations in acute myeloid leukemia. Blood 1999; 93: 3074–3080. 10 Rombouts WJ, Blokland I, Lowenberg B, Ploemacher RE. Biological characteristics and prognosis of adult acute myeloid leukemia with internal tandem duplications in the Flt3 gene. Leukemia 2000; 14: 675–683. 11 Kottaridis PD, Gale RE, Frew ME, Harrison G, Langabeer SE, Belton AA et al. The presence of a FLT3 internal tandem duplication in patients with acute myeloid leukemia (AML) adds important prognostic information to cytogenetic risk group and response to the first cycle of chemotherapy: analysis of 854 patients from the United Kingdom Medical Research Council AML 10 and 12 trials. Blood 2001; 98: 1752–1759. 12 Moreno I, Martin G, Bolufer P, Barragan E, Rueda E, Roman J et al. Incidence and prognostic value of FLT3 internal tandem duplication and D835 mutations in acute myeloid leukemia. Haematologica 2003; 88: 19–24. 13 Iwai T, Yokota S, Nakao M, Okamoto T, Taniwaki M, Onodera N et al. Internal tandem duplication of the FLT3 gene and clinical evaluation in childhood acute myeloid leukemia. The Children’s Cancer and Leukemia Study Group, Japan. Leukemia 1999; 13: 38–43. 14 Meshinchi S, Woods WG, Stirewalt DL, Sweetser DA, Buckley JD, Tjoa TK et al. Prevalence and prognostic significance of Flt3 internal tandem duplication in pediatric acute myeloid leukemia. Blood 2001; 97: 89–94. 15 Whitman SP, Archer KJ, Feng L, Baldus C, Becknell B, Carlson BD et al. Absence of the wild-type allele predicts poor prognosis in adult de novo acute myeloid leukemia with normal cytogenetics and the internal tandem duplication of FLT3: a cancer and leukemia group B study. Cancer Res 2001; 61: 7233–7239. 16 Stirewalt DL, Kopecky KJ, Meshinchi S, Appelbaum FR, Slovak ML, Willman CL et al. FLT3, RAS, and TP53 mutations in elderly patients with acute myeloid leukemia. Blood 2001; 97: 3589–3595. 17 Kussick SJ, Wood BL. Using four-color flow cytometry to identify abnormal myeloid populations. Arch Pathol Lab Med 2003; 127: 1140–1147. 18 Stirewalt DL, Meshinchi S, Kussick SJ, Sheets KM, PogosovaAgadjanyan E, Willman CL et al. Novel FLT3 point mutations within exon 14 found in patients with acute myeloid leukaemia. Br J Hematol 2004; 124: 481–484. 19 Seabright M. A rapid banding technique for human chromosomes. Lancet 1971; 2: 971–972. Leukemia


1598 20 Radich JP, Gehly G, Gooley T, Bryant E, Clift RA, Collins S et al. Polymerase chain reaction detection of the BCR-ABL fusion transcript after allogeneic marrow transplantation for chronic myeloid leukemia: results and implications in 346 patients. Blood 1995; 85: 2632–2638. 21 Borrow J, Goddard AD, Gibbons B, Katz F, Swirsky D, Fioretos T et al. Diagnosis of acute promyelocytic leukaemia by RT-PCR: detection of PML-RARA and RARA-PML fusion transcripts. Br J Haematol 1992; 82: 529–540. 22 Wetzler M, McElwain BK, Stewart CC, Blumenson L, Mortazavi A, Ford LA et al. HLA-DR antigen-negative acute myeloid leukemia. Leukemia 2003; 17: 707–715. 23 Munoz L, Aventin A, Villamor N, Junca J, Acebedo G, Domingo A et al. Immunophenotypic findings in acute myeloid leukemia with FLT3 internal tandem duplication. Haematologica 2003; 88: 637–645.

Leukemia

24 Noguera NI, Breccia M, Divona M, Diverio D, Costa V, De Santis S et al. Alterations of the FLT3 gene in acute promyelocytic leukemia: association with diagnostic characteristics and analysis of clinical outcome in patients treated with the Italian AIDA protocol. Leukemia 2002; 16: 2185–2189. 25 Arrigoni P, Beretta C, Silvestri D, Rossi V, Rizzari C, Valsecchi MG et al. FLT3 internal tandem duplication in childhood acute myeloid leukaemia: association with hyperleucocytosis in acute promyelocytic leukaemia. Br J Haematol 2003; 120: 89–92. 26 Shih LY, Kuo MC, Liang DC, Huang CF, Lin TL, Wu JH et al. Internal tandem duplication and Asp835 mutations of the FMS-like tyrosine kinase 3 (FLT3) gene in acute promyelocytic leukemia. Cancer 2003; 98: 1206–1216.