¹Department of Hematology and Oncology, University of Göttingen, Göttingen, Germany

²Medzinische Klinik III, University of Munich, Klinikum Grohadern, Munich, Germany

³Department of Hematology and Oncology, University of Münster, Münster, Germany

Correspondence to: S Schnittger, Labor für spezielle Leukämiediagnostik, Medzinische Klinik III der Universität München im Klinikum Grohadern, Marchioninistrae 15, 81377 München, Germany; Fax: 49 89 7095 4971

Partial tandem duplications of the MLL gene have been associated with trisomy 11 in acute myeloid leukemia (AML) and recently, have also been reported for karyotypically normal AML. In order to test the incidence and prognostic importance of this molecular marker, we have analyzed eight cases of AML with trisomy 11 and 387 unselected consecutive cases with AML for partial duplications of the MLL gene. Patients with normal karyotypes and those with various chromosome aberrations were included. De novo as well as secondary leukemias including all FAB subtypes were analyzed. Performing a one-step RT-PCR with 35 cycles using an exon 9 forward primer and an exon 3 reverse primer partial tandem duplications of the MLL gene were demonstrated in 3/8 (37.5%) patients with trisomy 11. In addition, 13/387 (3.4%) of unselected cases revealed a tandem duplication. Ten of these 13 cases were cytogenetically normal, the other three cases had 2 additional chromosomal alterations. Sequencing of the RT-PCR products of all 16 positive cases revealed fusions of MLL exon 9/exon 3 (e9/e3) (six cases), e10/e3 (three cases), e11/e3 (four cases) or combinations of differentially spliced e10/e3 and e11/e3 (three cases) transcripts. The duplications were confirmed by genomic long range PCR and Southern blot hybridization. Twelve cases with the MLL duplication were de novo myeloid leukemia, one was a secondary AML after MDS, three were therapy-related AML (t-AML). Of the 16 MLL-duplication positive cases, seven were classified as FAB M2, two as M1, five as M4, one as M0, one as M5b. The mean age was 62.3 years for patients with MLL duplication vs 50.3 years for the control group. Of 15 adult patients, 12 received treatment. Of these, three were non-responders, five had early relapse (6 months), four relapsed between 7 and 12 months. Median survival and relapse-free interval of the MLL duplication positive group was significantly worse than those of an age-matched karyotypically normal control group. In conclusion, MLL tandem duplications (1) are less common than previously reported; (2) are preferentially observed in AML with normal karyotypes, but can also be found in the presence of chromosome alterations; (3) are not strongly associated with an FAB subtype; (4) were not observed with the prognostically favorable t(8;21), inv(16), and t(15;17), other recurrent translocations, or in complex karyotypes; and (5) identifies a subgroup of patients with an unfavorable prognosis. Leukemia (2000) 14, 796-804.

MLL gene; tandem duplication; AML; trisomy 11

Introduction

The characterization of chromosome alterations in acute myeloid leukemia (AML) has allowed the establishment of biological and prognostic subgroups. However, nearly 50% of AML cases reveal normal karyotypes and lack molecularly detectable genetic markers. Thus, new molecular markers for subclassification of AML and as PCR targets for follow-up studies are eagerly awaited.

Translocations involving 11q23 with more than 30 different chromosomal sites resulting in MLL fusion genes (MLL for mixed lineage leukemia) have been described in ALL, as well as in 5-10% of AML (for review see Ref. ¹). Rearrangements with the different partner genes are strikingly diverse, but all affect a common breakpoint area of the MLL gene extending 8.3 kb between exons 5 (exon 8 according to the new nomenclature²) and exon 9 (12). Typically, AML patients with 11q23 translocations have acute myelomonocytic leukemia or acute monoblastic leukemia, and often have biphenotypic antigene expression. Clinically, adult patients with 11q23 translocations have a short disease-free survival and a poor prognosis.^3,4 However t(9;11) in childhood and adult AML may be prognostically favorable.^5,6,7,8,9 Hence, identification of patients with MLL gene rearrangements could have important implications with regard to treatment. Recently, a direct partial tandem duplication within the MLL gene has been described.^10,11,12 This rearrangement leads to fusion of a portion of the putative proto-oncogene with itself, and this seems to represent a new genetic mechanism for leukemogenesis.

The MLL duplication was first described in AML with trisomy 11 as a sole abnormality. This is an infrequent nonrandom chromosomal aberration observed in AML or myelodysplastic syndrome (MDS). More than 50 cases have been reported,^{13,14,15,16,17,18} but trisomy 11 does not correlate with any specific subtype of the FAB classification. AML with trisomy 11 was shown to be associated with a stem/progenitor cell immunophenotype with myeloid antigen expression and is characterized by poor response to standard chemotherapy and an unfavorable prognosis.¹⁸ Recently, molecular characterization of de novo AML with trisomy 11 has revealed that the MLL gene was rearranged without cytogenetic evidence of 11q23 translocations. This rearrangement was shown to represent a partial tandem duplication of the MLL gene.^{10,11,12,19,20} In most cases the duplicated region of the MLL gene spanned exons 2-6 (exons 3-9 according to Ref. ²) or exons 2-8 (exons 3-11, accordingly).

The MLL duplication has been found to be strongly associated with trisomy 11.^10,19 Caligiuri et al¹⁰ detected the MLL duplication in three of four patients with +11 as sole chromosomal abnormality. It was also found in 2/19¹⁰ and in 8/95²¹ karyotypically normal AML, and thus was thought to be a very promising new molecular marker for AML studies. In addition to these cases, about 20 more patients with AML showing MLL duplications have been described by different authors. Most of the published studies focused on special subgroups of leukemia and patients were selected based on specific FAB subtypes (French-American-British classification) or on normal karyotype. Thus the real frequency of MLL duplications in unselected AML cases remains unclear. In AML patients with normal karyotype a frequency of the MLL duplication of nearly 10% (2/19)²² up to 21% (7/34) has been suggested²³ and thus may prove to have a similar incidence as MLL fusions with partner genes, which have been estimated as 5% in adult AML.²⁴ Thus the aims of this study were: (1) to analyze the frequency of MLL duplications in 387 unselected cases of AML; (2) to correlate MLL duplication with karyotype, FAB subtype and age; (3) to investigate the outcome of patients with MLL duplications.

Materials and methods

Patient samples

Within a 2-year period from November 1996 to October 1998, fresh blood or bone marrow samples from 387 consecutive patients were analyzed. All cases were diagnosed as having AML according to standard French-American-British (FAB) classification^25,26 and were referred to our laboratory for cytogenetic analysis. With the exception of one child all patients were adults. Do novo as well as secondary AML after MDS or treatment of a previous malignant disease (t-AML) were included. Additionally, eight patients with a known trisomy 11 (Table 1) were retrospectively included in the analysis.

Treatment protocol

Six patients (cases 4, 7, 8, 10, 14 and 16) were treated according to the AMLCG92 protocol as published.²⁷ Treatment consisted of TAD9 (6-thioguanine, cytosine-arabinoside, daunorubicin) induction (age >60 years) or TAD9/HAM (6-thioguanine, cytosine-arabinoside, daunorubicin/high-dose cytosine-arabinoside, mitoxantrone) double induction (age <60 years) and TAD9 consolidation therapy. Six patients (cases 1, 2, 3, 6, 12, 15) were not included within the AMLCG92 study but were treated with analogous protocols or even with more intensive chemotherapy (HAM) (cases 3 and 12). Patients were either put on maintenance therapy or received S-HAM (sequential HAM) for late consolidation (case 4). Three adult patients (5, 9, 13) did not receive chemotherapy due to bad performance status, one child (case 11) was treated according to the German pediatric AML protocol and subsequently underwent allogeneic bone marrow transplantation.

Cytogenetics

Cytogenetic G-banding analysis was performed with standard methods. The definition of a cytogenetic clone and descriptions of karyotypes followed the International System for Human Cytogenetic Nomenclature.²⁸

Nucleic acid isolation

DNA was extracted with a salting out procedure²⁹ from fresh bone marrow or peripheral blood cells after Ficoll separation of mononucleated cells. From the same specimens total RNA was isolated as previously described.³⁰

RT-PCR

Reverse transcription and PCR was performed as previously described.³⁰ Primers for simple one-step PCR were: 6.1: 5'GTCCAGAGCAGAGCAAACAG3' (bp 4013-4032,¹⁹ numbering according to Ref. ³¹) and E3AS: 5'ACACAGATGGA TCTGAGAGG3' (bp 567-586).²³ For nested PCR primers 3.C1 5'AGGAGAGAGTTTACCTGCTC3' (bp 821-840),¹¹ and MLLint. 5'CTTCCAGGAAGTCAAG CAAGCAGGT3' (bp 3869-3892) were used in the primary reaction and a 1 l aliquot of these reactions was further amplified for an additional 35 cycles with primers 6.1 and E3AS. The positions of primers within the MLL gene is indicated in Figure 1.

For each RNA sample an ABL specific RT-PCR was performed to control the integrity of RNA using primers abl5': 5'GGCCAGTAGCATCTGACTTTG3' and abl3': 5'ATGGTACCAGGAGTGTTTCTCC3'. Water instead of cDNA was included as a blank sample in each experiment.

Sensitivity of PCR

One microgram of RNA was isolated from about 100 000 cells and was reverse transcribed in one experiment. As we used 1/40 l of the RT reaction for PCR, we analyzed an equivalent of 2500 cells per sample.

Dilution experiments with cDNA from a duplication positive AML and a normal donor sample were performed in 10^-1 steps up to 10^-8. To equalize the total amount of cDNA in the reaction mixture cDNA prepared from mouse spleen was added. Mouse spleen RNA was previously found to be negative for MLL-duplication transcripts.³⁰

Sequencing

All RT-PCR products were sequenced as previously described³¹ to identify the involved exons and the extent of the duplication.

Nomenclature

We used the new MLL exon nomenclature according to Nilson et al² that differs from the widespread older nomenclature (see Figure 1).

Genomic XL-PCR

Long range genomic amplification (XL-PCR) with primers from exon 9: 6.1 (5'GTCCAGAGCAGAGCAAACAG3')¹⁹ and exon 3: 2.0 (5'CGCACTCTGACTTCTTCATC3')¹⁹ was performed with 100 ng DNA using the XL-PCR kit (Perkin Elmer, Weiterstadt, Germany) following the manufacturer's instructions. Cycling conditions were 1 min 94°C, 10 min 65°C with 15 s extension for 35 cycles. As a control for DNA integrity and for the XL-PCR conditions the wild-type allele was amplified as a control with primers E5S1 (5'GAGAGGATCCTGCCC- CAAAG3') and E11AS (5'CATGAGACCACTGTGCATCC).²³

Southern blot analysis

DNA from all PCR-positive patients was re-evaluated by Southern blot analysis. In addition, some duplication negative patients were included. Each 5-10 g DNA was digested with the restriction enzymes BamHI and EcoRI overnight. Completely digested DNA was size fractionated on 0.7% agarose gels and transferred to nylon membranes (SureBlot, Oncor, Heidelberg, Germany). A diogixin-dUTP labelled MLL cDNA probe (Oncor) containing exons 8-15 was used for hybrization. The membranes were washed under high stringency (0.1 ´ SSC for 2 ´ 15 min at 65°C). Detection was performed with an anti-digoxin antibody (Boehringer, Ingelheim, Germany) and CSPD (Boehringer).

Statistical analysis

For each patient with an MLL duplication, two control patients were identified. Identification criteria were solely age and normal karyotype. A deviation of age ±1 year for the control group was permitted. Survival curves were calculated for overall survival and relapse-free interval according to Kaplan-Meier.³² Survival as well as relapse-free interval were calculated from the start of therapy. Patients dying in CR were censored in the analysis of relapse-free interval. Survival curves were compared using a double sided log-rank-test, results were significant at a level of P < 0.05 at both sides.

Results

Incidence of MLL duplication in 395 AML patients

Three hundred and ninety-five AML patients were analyzed for the presence of a MLL duplication. The patients were divided into two subgroups. (1) Eight cases were retrospectively analyzed because of a trisomy 11 (two with +11 as sole karyotypic aberration, three with +11 and one or two additional chromosomal changes and three with +11 and complex karyotypes). MLL duplication was found in three of these patients (37.5%). One had +11 as sole karyotypic abnormality (case 1), one had del(13)(q12q14) (case 2) and one del(5)(q15q33) (case 3) in addition to trisomy 11 (Table 1). Sequence analysis revealed an exon 10/exon 3 fusion (case 1), an exon 11/exon 3 fusion in cases 2, and case 3 had alternatively spliced exon 10/exon 3 and exon 11/exon 3 fusions (Table 1). In addition we analyzed two cases with i(11q) but in contrast to Caligiuri et al¹⁹ who observed a duplication in 1/3 cases with +11q, our two cases with +11q were negative for the duplication. (2) 387 consecutive cases were prospectively analyzed. All FAB subtypes were included, and no selection was performed according to de novo or secondary leukemia or karyotype. The primary screening was performed using RT-PCR. Amplification products of different sizes were found in 13 patients (3.36%) (Figure 2; Table 1). The specificity of each PCR fragment was confirmed by nucleotide sequencing. Fusions of exon9/exon3 were found in six cases, of exon10/exon3 in two cases, of exon 11/exon 3 in two cases. In addition, three cases revealed two different fragments corresponding to alternatively spliced exon 10 and 11 to exon 3 fusion transcripts (Table 1, Figure 2). Long range PCR revealed only one amplification product in each of these cases (Figure 4) indicating that the different transcription fragments are the result of a differential splicing and not of two differently altered alleles.

In total 3.4% (13/387) of the unselected AML patients were found to have a MLL tandem duplication. Ten of these had a normal karyotype (cases 4-9, 12-14, 16). Three cases revealed chromosome aberrations: a del 7(q22) (case 11), a t(1;16)(q21;q22) (case 10), and a del(9)(q22) (case 15), respectively (Table 1). Thus, MLL duplications are not restricted to karyotypically normal AML.

Confirmation of RT-PCR data by Southern blotting

All RT-PCR positive cases were re-evaluated by Southern blot analyses using two different restriction enzymes (Figure 3). One rearranged fragment was demonstrated after restriction enzyme digestion with BamHI or EcoRI in all cases in addition to the germline band indicating a rearrangement within the MLL gene. These data are consistent with a monoallelic rearrangement and exclude MLL duplication as a result of non-homologous recombination between both MLL alleles that would result in a MLL deletion in the second allele. The signal intensity corresponded approximately to the blast count of the analyzed samples. MLL duplication negative cases only revealed the germline band.

MLL duplications as marker for follow-up studies

Cases 3 and 4 relapsing at 7 months and 2 months were monitored by RT-PCR for the MLL duplication. Relapse samples showed the same MLL exon fusions as at diagnosis (Figure 5 and Table 1). Cytogenetics revealed a normal karyotype at diagnosis and gains of chromosomes 8, 13 and 19 at relapse. Case 3 revealed additional structural alterations presenting as del(17)(q23) and add(7)(q3?4) at relapse 7 months after diagnosis. Thus, MLL duplication may represent an early event in the malignant transformation process, which seems to be a stable molecular marker and may therefore be helpful to distinguish true relapse from a secondary leukemia.

Performing nested PCR with a sensitivity of 1/2500 cells we detected MLL duplication transcripts in all AML samples (Figure 6) corresponding to the data we have described for normal individuals.³⁰ Both AML-typical as well as transcripts different from those detected in AML (exon 9/3, 10/3, and 11/3 fusions) were observed. Cases revealing duplication transcripts only after nested RT-PCR, do not show rearranged bands by Southern blot analysis due to the few cells that carry the duplication. From these data we can conclude that the only reliable methods for the diagnosis of the AML related MLL duplication are one-step RT-PCR, genomic PCR and Southern blot analysis. In addition, these observations limit the reliability of this molecular marker for the detection of minimal residual disease in follow-up studies.

Association of MLL duplication positive cases with age

Of the 387 cases, complete data on age, FAB subtypes, and cytogenetics are available in 217 cases. The mean age of the control cohort was 50.3 years. The mean age of the duplication positive cases excluding the only child (case 11) was 62.3 years. This difference was statistically not significant.

Correlation of MLL duplication positive cases with FAB subtype

Within the control cohort of 217 cases, six cases were AML M0 (2.8%), 32 AML M1 (14.7%), 60 AML M2 (27.6%), 16 AML M3 (7.4%), 40 AML M4 (18.4%), 20 AML M4eo (9.2%), 21 AML M5 (9.7%), seven AML M6 (3.2%), two M7 (0.9%), and 13 cases (6%) were not classified (Table 3). Twenty-six cases (12%) were known to have s-AML after MDS or treatment for a previous malignant disease (t-AML) and 10(4.6%) were AML relapses.

The tandem duplication was present in different FAB subtypes. Of the 16 cases, one was M0 (6.3%), two M1 (12.5%), seven M2 (43.8%), five M4 (31.25%), one M5b (6.3%) (Table 1). Thus, the majority of MLL duplication positive cases were observed in M2, reflecting the normal FAB subtype distribution in the control cohort.

Occurrence of MLL duplication in therapy associated and AML secondary to MDS

Twelve cases positive for the MLL duplication had de novo leukemia, one patient had secondary AML after MDS prephase (case 15), and three patients had treatment related AML: case 11 after T-ALL, case 13 after AML, and case 16 after bladder and prostate cancer. Thus 4/16 (25%) had secondary AML vs 26/217 (12%) in the control cohort.

Incidence of MLL duplication in cytogenetically defined subgroups

Of the control cohort of 217 cases, 105 (49.4%) were karyotypically normal and 112 (51.6%) showed various common or uncommon chromosome aberrations. Forty-six (21.2%) of all cases had rearrangements like inv(16) (20 cases), t(15;17) (16 cases), t(8;21) (10 cases) typically associated with AML.

Within this cohort of 217 adult cases, seven cases (4-10, 12) were MLL duplication positive (3.2%). The percentage of positive cases is slightly higher with 4.3% (7/164) if AML with typical rearrangements like inv(16), t(15;17), t(8;21), t(3;21), and t(6;9) were excluded. Fifty-two of all patients had one of these aberrations and none of those cases was positive for the MLL duplication. If only AML with normal karyotype are regarded, 5.7% (6/105) were positive for the duplication.

As MLL duplications were associated with FAB M2 subtype in four of seven unselected cases (cases 4-7) analyzed within the control cohort of 217 cases, the incidence of MLL duplication positive AML-M2 without t(8;21) was four of 51 (7.8%), thus the incidence of MLL duplication was highest in the group of FAB M2 without t(8;21).

Outcome of adult patients with MLL duplication positive AML

Of the 15 adult patients with MLL duplication, three patients did not receive treatment (Table 2). Of the 12 patients receiving therapy, three did not achieve complete remission after the first TAD9 induction course. Of the nine remitters, two did not get any further therapy, one received an additional course of TAD9, six patients received one or two high-dose Ara-C/mitoxantrone courses. Of the nine responders, eight have relapsed: four patients with early relapse at 2, 4, 4, and 5 months (cases 2, 6, 10, and 14, respectively) and four patients with later relapses at 8, 8, 9, 11 months (cases 3, 4, 15, 12, respectively). One patient died in aplasia with complete blast clearance in the bone marrow (case 8). The only childhood case achieved CR and died of complications of allogeneic bone marrow transplantation at 10 months.

The median survival of all cases was 5 months, the median relapse-free interval of responders was 4 months (Figure 7). A matched pair analysis with two karyotypically normal, age matched control cases for each MLL duplication patient was performed. Control cases were treated within the AMLCG92 study and received TAD9 induction (>60 years) or TAD9-HAM induction (<60 years) and TAD consolidation. Patients then were randomized to receive either 24 cycles of maintenance therapy or intensified consolidation with S-HAM. The median survival of the control group was 12 months, the median relapse-free interval of responding patients within the control group was not reached, the Kaplan-Meier projection of relapse-free rate is 53.6% (Figure 8). Differences were statistically significant with a P value of 0.00015 for relapse-free interval and 0.0060 for survival (double sided log-rank test).

Discussion

The frequencies of MLL duplications in AML with trisomy 11, as well as in unselected AML was below the incidences reported in previous publications (Table 4). The frequency of duplication-positive cases in patients with trisomy 11 was 37.5% (3/8). Frequencies of MLL duplications of 79% in patients with +11 as sole karyotypic abnormality³³ and of 91% of patients with +11 with and without further alterations¹⁹ were reported. As two of our cases revealed additional aberrations, such as del(13)(q12q14) (case 2) and del(5)(q15q33) (case 3), we confirm that MLL duplications occur not only in cases with +11 as sole karyotypic aberration but also in +11 AML with one or two additional chromosomal alterrations. In agreement with this report,¹⁹ no duplication-positive cases were found in cases with +11 and complex karyotype (three patients analyzed). Restricting our analysis to +11 cases without complex karyotype the frequency in our cohort rises to 60% (3/5). The MLL tandem duplication is the first molecular change in leukemia associated with a chromosome gain. The molecular mechanism leading to the MLL duplication is still unresolved and it is completely unclear why it occurs preferentially in addition/or together with gain of chromosome 11.

The frequency of MLL duplications (3.4%) in unselected AML is also below the frequency of previous studies (Table 4). Some explanations may be: (1) regional or genetic population differences as have been reported for the t(15;17)/PML-RAR can not be ruled out. (2) Different sensitivity levels in the different studies. As previously reported,³⁰ MLL duplication transcripts are observed in virtually all healthy individuals at a level of 1/250-1000 cells. Sensitivity of RT-PCR in our screening was set in a way that fusion transcripts in healthy donors were not detected by a one-step RT-PCR with 35 cycles. In addition, all RT-PCR results were confirmed by genomic XL-PCR and Southern blot analyses. From intensity of rearranged hybridization fragments of Southern blots we conclude that at least 20% of cells carry the duplication. Other studies did not quantify the amount of duplication positive cells. Yu et al²³ described multiple duplication transcripts including also minor transcripts with out of frame fusions. It is suggestive that some of these fusions may represent 'contaminating' duplication transcripts of normal hematopoietic cells.³⁰ (3) In most of the previous studies, only restricted AML types, such as karyotypically normal AML or AML FAB M2 subtypes, were analyzed. In contrast, we have analyzed unselected AML cases. If frequencies in our analysis are calculated only for karyotypically normal cases (5.7%) or for AML FAB M2 including t(8;21) (6.7%) or exluding t(8;21) (7.8%), they approach the incidences reported by other groups. In addition, our analysis included AML patients irrespective of treatment protocol. Thus our cohort may be more representative for the biological spectrum of AML. A similar incidence of <10% was reported by Caligiuri et al¹⁰ who included secondary leukemias and cases with various cytogenetic alterations in their screening and detected 2/33 (6.1%) MLL duplication-positive cases without trisomy 11.

In accordance with previous studies the MLL duplication was not found to be restricted to a specific FAB subtype, but there was a tendency towards a higher incidence within the M2 subtype, as 7/16 (43.8%) duplication-positive cases were M2, whereas only 60/217 (27.6%) in the control group were M2 (Table 3). This is in contrast to translocations involving MLL which predominantly occur in the myelomonocytic (M4) or monocytic (M5) subtypes. As the MLL duplication has been found in M0, M1, M2, M4, M5 (own data) and M6,³⁴ the molecular event does not seem to be restricted to a specific stage of differentiation or lineage commitment within the myeloid compartment.

MLL duplications were observed in cases with relatively frequent cytogenetic markers, such as del(7)(q22) (case 11), del(9)(q22) (case 15), and together with rare aberrations (case 10). Whether these alterations are primary or secondary to the MLL duplication is unclear. However, two patients revealed structural alterations, such as del(17)(q23) and add(7)(q3?4) (case 3) or gains of chromosomes 8, 13, and 19 (case 4) at relapse in addition to the MLL duplication that was already present at diagnosis. The type of duplication fusion transcripts remained constant in both cases, suggesting that MLL duplication may represent an early, if not the initiating event, in the leukemic transformation.

MLL duplications were not observed in association with frequent and prognostically favorable alterations, such as t(15;17), t(8;21), or inv(16), nor were MLL duplications found in the prognostically unfavorable subgroup with complex karyotypes, suggesting a certain mechanism for transformation in MLL duplication-positive leukemia.

In our cohort of 16 duplication-positive cases, 12 cases were presenting with de novo leukemia, and four with secondary AML: one after MDS prephase (case 15) and three with preceding chemotherapy (cases 11, 13, 16). Case 11 was diagnosed 4 years after treatment of T-ALL and an MDS prephase of 6 months. It has to be mentioned that this is the only child in our study and was analyzed by chance. Although MLL translocations are more common in children than in adults, no child with MLL duplication has been reported so far, and the frequency of MLL duplication in childhood is unclear. Case 13 was diagnosed as AML relapse 5 years after achieving CR, but was reclassified as secondary leukemia, as no MLL rearrangement was detected in the material of the initial AML by Southern blotting. With the exception of one case with MDS-derived overt leukemia with trisomy 11 and exon 8/2 fusion,³³ no further secondary leukemia with MLL duplication has been reported. Our data support that MLL duplications do occur in therapy-related AML, as has been shown for translocations involving the MLL locus. No obvious differences between the MLL duplications in t-AML and de novo AML were observed at the cDNA and the DNA level by RT-PCR and XL-PCR, respectively. Therefore, it is likely that ALU recombination also mediates MLL duplication in t-AML as has been described for de novo AML.^1,11,35,36

MLL duplications may represent a potential unique molecular marker for a subset of AML cases which are otherwise not amenable for detection of minimal residual disease by cytogenetics of FISH. However, as published,³⁰ MLL duplications frequently occur in normal hematopoiesis and are easily detectable by nested RT-PCR. The biological significance of this finding is still unclear, but renders nested RT-PCR for MLL duplications unsuitable for MRD monitoring. Whether quantitative 'real-time' PCR for the patient specific fusion transcript may overcome this problem, remains to be evaluated.

The prognostic relevance of complex karyotypes as well as of the 'favorable' t(8;21), inv(16) and t(15;17) have been established by several groups, although treatment variables as well as the still unresolved impact of additional chromosomal alterations still hamper definitive statements.^37,38 However, risk factors within the prognostically intermediate group with normal karyotype are lacking and eagerly awaited. Previous publications with limited numbers of patients have suggested that the presence of an MLL duplication may have an unfavorable impact on survival.^21,35 Yu et al²³ compared seven patients with MLL duplication with 27 patients with AML, which were not further defined. They found a median survival of 2.7 months of the MLL duplication group vs 6.8 months of the control group. Caligiuri et al²¹ compared seven patients with MLL duplication achieving complete remission (CR) with 61 control patients who had de novo AML and exhibited a normal karyotype. MLL duplication-positive cases had a median relapse-free interval of 7.1 months and the control group of 23.2 months. This difference was statistically significant. In our group of 16 patients, 15 adult patients were evaluated for outcome. Although the CR rate was high with nine of 12 treated patients achieving CR, the median survival was only 5 months and the median relapse-free interval of the nine responders was 4 months. Although these are still small numbers, they confirm that MLL duplication heralds a bad prognosis.

In our series, MLL duplication-positive cases were 12 years older than the control cohort. In order to rule out a potential impact of age upon prognosis, a matched pair analysis was performed establishing a control group within the AMLCG92 study.²⁷ For each MLL duplication-positive case, two cases matched for age and having a normal karyotype were selected. The median survival of the 15 adult patients (5 months) was statistically significantly lower than that of the matched pair control group of 30 patients (12 months). Also, the relapse-free interval of the MLL duplication-positive group (4 months) was significantly shorter than that of the control group, which had a proportion of relapse-free survival of 53.6%. All 12 treated patients with MLL duplications received at least one course of TAD9, six of nine responders received one to two courses of high-dose Ara-C (1 g/m² for patients >60 years of age, 3 g/m² for patients <60 years of age) and even in this small subgroup of patients, median relapse-free interval was only 7 months, and median survival 10 months, comparable to bad prognosis AML.³⁷ In conclusion, MLL duplication is associated with a bad prognosis and there is a suggestion that MLL duplication-positive AML may not be eradicated by high dose Ara-C. A new prognostically unfavorable AML subgroup has been identified by molecular analysis. Whether patients with MLL duplications may benefit from different therapy management has to be clarified.

Acknowledgements

This work was supported by a program grant of the Deutsche Forschungsgemeinschaft to FG and BW (SFB500, project A1).

1 Bernard OA, Berger R. Molecular basis of 11q23 rearrangements in hematopoietic malignant proliferations. Genes Chromos Cancer 1995; 13: 75-85, MEDLINE

2 Nilson I, Löchner K, Siegler G, Greil J, Beck J, Fey G, Marschalek R. Exon/intron structure of the human ALL-1 (MLL) gene involved in translocations to chromosomal region 11q23 and acute leukemias. Br J Hematol 1996; 93: 966-972,

3 Drexler HG, Thiel E, Ludwig WD. Review of the incidence and clinical relevance of myeloid antigen-positive acute lymphoblastic leukemia. Leukemia 1991; 5: 637-645, MEDLINE

4 Cuneo A, Michaux JL, Ferrant A, Van HL, Bosly A, Stul M, Dal CP, Vandenberghe E, Cassiman JJ, Negrini M et al. Correlation of cytogenetic patterns and clinicobiological features in adult acute myeloid leukemia expressing lymphoid markers. Blood 1992; 79: 720-727,

5 Kalwinsky DK, Raimondi SC, Schell MJ, Mirro JJ, Santana VM, Behm F, Dahl GV, Williams D. Prognostic importance of cytogenetic subgroups in de novo pediatric acute nonlymphocytic leukemia. J Clin Oncol 1990; 8: 75-83, MEDLINE

6 Martinez-Climent CJ, Lane NJ, Rubin CM, Morgan E, Johnstone HS, Mick R, Murphy SB, Vardiman JW, Larson RA, Le Beau M, Rowley JD. Clinical and prognostic significance of chromosomal abnormalities in childhood acute myeloid leukemia de novo. Leukemia 1995; 9: 95-101, MEDLINE

7 Fenaux P, Preudhomme C, Lai J, Morel P, Beuscart R, Bauters F. Cytogenetics and their prognostic value in de novo acute myeloid leukaemia: a report on 283 cases. Br J Haematol 1989; 73: 61-67, MEDLINE

8 Mrozek K, Heinonen K, Lawrence D, Carroll AJ, Koduru PR, Rao KW, Strout MP, Hutchison RE, Moore JO, Mayer RJ, Schiffer CA, Bloomfield CD. Adult patients with de novo acute myeloid leukemia and t(9;11)(p22;q23) have a superior outcome to patients with other translocations involving band 11q23: a Cancer and Leukemia Group B study. Blood 1997; 90: 4532-4538, MEDLINE

9 Baer MR, Stewart CC, Lawrence D, Arthur DC, Mrozek K, Strout MP, Davey FR, Schiffer CA, Bloomfield CD. Acute myeloid leukemia with 11q23 translocations: myelomonocytic immunophenotype by multiparameter flow cytometry. Leukemia 1998; 12: 317-325, MEDLINE

10 Caligiuri M, Schichman SA, Strout MP, Mrozek K, Baer MR, Frankel SR, Barcos M, Herzig GP, Croce CM, Bloomfield CD. Molecular rearrangement of the ALL-1 gene in acute myeloid leukemia without cytogenetic evidence of 11q23 chromosomal translocations. Cancer Res 1994; 54: 370-373,

11 Schichman SA, Caligiuri M, Strout M, Carter S, Gu Y, Canaani E, Bloomfield C. Croce C. ALL-1 tandem duplication in acute myeloid leukemia with a normal karyotype involves homologous recombination between ALU elements. Cancer Res 1994; 54: 4277-4280, MEDLINE

12 Schichman S, Caligiuri M, Gu Y, Strout M, Canaani E, Bloomfield C, Croce C. ALL-1 partial duplication in acute leukemia. PNAS 1994; 91: 6236-6239,

13 Takasaki N, Kaneko Y, Maseki N, Sakurai M. Trisomy 11 in chronic myelomonocytic leukemia: report of two cases and review of the literature. Cancer Genet Cytogenet 1988; 30: 109-117,

14 Weh HJ, Hoffmann R, Suciu S, Kuse R, Hossfeld DK. Is trisomy 11 another nonrandom chromosomal anomaly in acute nonlymphocytic leukemia and myelodysplastic syndromes? Cancer Genet Cytogenet 1988; 35: 205-211, MEDLINE

15 Ohyashiki K, Nagasu M, Hojo H, Iwabuchi A, Ohyashiki JH, Toyama K. Myelodysplastic syndrome with trisomy 11 associated with polycythemia vera. Am J Hematol 1989; 31: 122-125,

16 Avanzi GC, Giovinazzo B, Rosso A, Depaoli L, Bertini M, Resegotti L, Pegoraro L. Trisomy 11 in myelodysplastic syndrome-derived acute myeloblastic leukaemias. Eur J Haematol 1989; 43: 173-177,

17 Bilhou-Babera C, Lesesve J, Marit G, Lafarge M, Destugue N, Goullin B, Arnoulet C, Stoppa A-M, Huguet F, Attal M, Lacombe F, Broustet A, Reiffers J, Bernard P. Trisomy 11 in acute myeloid leukemia: ten cases. Leukemia 1994; 8: 2240-2231,

18 Slovak ML, Traweek ST, Willman CL, Head DR, Kopecky KJ, Magenis RE, Appelbaum FR, Forman SJ. Trisomy 11: an association with stem/progenitor cell immunophenotype. Br J Haematol 1995; 90: 266-273, MEDLINE

19 Caligiuri MA, Strout MP, Schichman SA, Mrozek K, Arthur DC, Herzig GP, Baer MR, Schiffer CA, Heinonen K, Knuutila S, Nousiainen T, Ruutu T, Block AW, Schulman P, Pedersen BJ, Croce CM, Bloomfield CD. Partial tandem duplication of ALL1 as a recurrent molecular defect in acute myeloid leukemia with trisomy 11. Cancer Res 1996; 56: 1418-1425, MEDLINE

20 Bernard OA, Romana SP, Schichman SA, Mauchauffe M, Jonveaux P, Berger R. Partial duplication of HRX in acute leukemia with trisomy 11. Leukemia 1995; 9: 1487-1490,

21 Caligiuri MA, Strout MP, Lawrence D, Arthur DC, Baer MR, Yu F, Knuutila S, Mrozek K, Oberkircher AR, Marcucci G, de la Chapelle A, Elonen E, Block AW, Rao PN, Herzig GP, Powell BL, Ruutu T, Schiffer CA, Bloomfield CD. Rearrangement of ALL1 (MLL) in acute myeloid leukemia with normal cytogenetics. Cancer Res 1998; 58: 55-59, MEDLINE

22 Schichman S, Canaani E, Croce C. Self-fusion of the ALL1 gene. A new genetic mechanism for acute leukemia. JAMA 1995; 273: 571-576, MEDLINE

23 Yu M, Honoki K, Andersen J, Paietta E, Nam DK, Yunis JJ. MLL tandem duplication and multiple splicing in adult acute myeloid leukemia with normal karyotype. Leukemia 1996; 10: 774-780, MEDLINE

24 Rowley JD. Rearrangements involving chromosome band 11q23 in acute leukaemia. Semin Cancer Biol 1993; 4: 377-385, MEDLINE

25 Bennett J, Catovsky D, Daniel M-T, Flandrin G, Galton D, Gralnick H, Sultan C. Proposals for the classification of acute leukemias, a report of the French-American-British Cooperative Group. Br J Haematol 1976; 33: 451-458, MEDLINE

26 Bennett J, Catovsky D, Daniel M-T, Flandrin G, Galton D, Gralnick H, Sultan C. Criteria for the diagnosis of acute leukemias of megakaryocyte lineage (M7), a report of the French-American-British Cooperative Group. Ann Intern Med 1985; 103: 460-462, MEDLINE

27 Büchner T, Hiddemann W, Wörmann B. Intensive consolidation versus prolonged maintenance following intensive induction and conventional consolidation in primary AML: a study by AMLCG. Blood 1996; 88: (Suppl 1) 214A,

28 ISCN. Guidelines for Cancer Cytogenetics. Karger: Basel, 1995,

29 Miller S. A simple salting out procedure for extracting DNA from nucleated cells. Nucleic Acid Res 1988; 16: 1215-1216, MEDLINE

30 Schnittger S, Wörmann B, Hiddemann W, Griesinger F. Partial tandem duplications of the MLL gene are detectable in peripheral blood and bone marrow of nearly all healthy donors. Blood 1998; 92: 1728-1734, MEDLINE

31 Gu Y, Nakamura T, Alder H, Prasad R, Canaani O, Cimino G, Croce C, Canaani E. The t(4;11) chromosome translocation of human acute leukemias fuses the ALL-1 gene, related to drosophila trithorax to the AF-4 gene. Cell 1992; 71: 701-708, MEDLINE

32 Kaplan E, Meier P. Nonparametric estimations from incomplete observations. J Am Stat Assoc 1958; 53: 457-481,

33 Yamamoto K, Hamaguchi H, Nagata K, Kobaqashi M, Taniwaki M. Tandem duplication of the MLL gene in myelodysplastic syndrome-derived overt leukemia with trisomy 11. Am J Hematol 1997; 55: 41-45,

34 Kwong Y. Partial duplication of the MLL gene in acute myelogenous leukemia without karyotypic aberration. Cancer Genet Cytogenet 1997; 97: 20-24,

35 So CW, Ma ZG, Price CM, Dong S, Chen SJ, Gu LJ, So CK, Wiedemann LM, Chan LC. MLL self fusion mediated by Alu repeat homologous recombination and prognosis of AML-M4/M5 subtypes. Cancer Res 1997; 57: 117-122,

36 Strout M, Marcucci G, Bloomfield C, Caligiuri M. The partial tandem duplication of ALL1 (MLL) is consistently generated by Alu-mediated homologous recombination in acute myeloid leukemia. Proc Natl Acad Sci USA 1998; 95: 2390-2395, Article MEDLINE

37 Bloomfield C, Lawrence D, Byrd J, Carroll A, Pettenati M, Tantravahi R, Patil S, Davey F, Berg D, Schiffer C, Arthur D, Mayer R. Frequency of prolonged remission duration after high-dose cytarabine intensification in acute myeloid leukemia varies by cytogenetic subtype. Cancer Res 1998; 58: 4173-4179, MEDLINE

38 Schlenk R, Fischer K, Del Valle F, Harmann F, Pralle H, Fischer J, Gunzer U, Pezzutto A, Weber W, Grimminger W, Preiss J, Göckel F, Haase R, Dohner H. Stratification of postremission therapy in adult acute myeloid leukemia according to the karyotype - preliminary results of a multicenter treatment trial AML HD93. In: Hiddemann W, Büchner T, Wörmann B, Ritter J, Creutzig U, Keating M, Plunkett W (eds). Acute Leukemias VII: Experimental Approaches and Novel Therapies. Springer: Berlin, 1998, 867-876.

Figure 1  Scheme of the 5' part of the MLL gene that contains the common breakpoint cluster region (bcr) of MLL translocations. We use the new exon nomenclature according Nilson et al.² The old nomenclature is given above. PCR primers used for detection of the duplications are indicated.

Figure 2  RT-PCR amplification of the MLL duplication transcripts presented on an ethidium bromide stained agarose gel. M: molecular weight standard VI (Boehringer); +C: sample with known MLL duplication; -C: blank control containing water instead of cDNA; P4-P10: patient numbers corresponding to the numbers in Table 1. Some sizes of the molecular weight standard are given on the left and kinds of the respective fusions are indicated at the right.

Figure 3  Representative Southern blot hybridization of BamHl digested DNA of AML patients hybridized with the exon 8-exon 15 probe. P: patient No. corresponding to No. in Table 1. C1-C4: AML patients without MLL rearrangement. Kind of exon fusions of the corresponding cases are given at the bottom of each lane. Estimated sizes of rearranged fragments in kb are given on the left hand side. The band of P9 is a double band.

Figure 4  Genomic long range (XL-) PCR of the MLL duplication. (a) Amplification of the wild-type MLL allele from exon 8 to exon 14 as a control PCR. (b) Amplification of the MLL duplication; sizes of the amplified fragments are indicated at the right; M1: molecular weight standard II (Boehringer); M2: molecular weight standard III (Boehringer); -C: blank control containing water instead of DNA, P2, P3, P4 are positive and P1, P5, P6, P7 are negative for MLL duplication.

Figure 5  RT-PCR follow-up of a patient with normal karyotype at diagnosis and with chromosome gains at relapse, showing that the exon 9/exon 3 fusion is stable at both time points, indicating that the MLL duplication is an early event in this AML. M: molecular weight standard VI (Boehringer), P9 and P3 were used as positive controls in this experiment; P4: case 4 at diagnosis, P4_R case 4 at relapse.

Figure 6  Nested RT-PCR of MLL duplication negative cases (A-K); corresponding to transcripts that can be found at a low level in healthy samples³⁰ and indicating that nested RT-PCR is not reliable for follow-up studies.

Figure 7  Kaplan-Meier plot of survival probability of 15 adult patients with MLL duplications compared with an age-matched control cohort with normal karyotypes. X-axis: survival in days; y-axis: percentage of survival. Median survival of the MLL duplication group was 5 months which was statistically significantly different to the control cohort with 12 months (P = 0.00600).

Figure 8  Kaplan-Meier plot of relapse-free interval of patients achieving CR estimated from the time point of CR: nine adult patients with MLL duplications are compared with 17 age-matched patients with normal karyotype. X-axis: relapse-free interval in days; y-axis: percentage of relapse-free interval. Median relapse-free interval of the MLL duplication group was 4 months, the median relapse-free interval of the control group has not been reached. This difference was statistically significantly different (P = 0.00015).

Table 1  Characterization of patients with MLL duplications

Table 2  Clinical data of patients with MLL duplication

Table 3  Frequency of MLL duplication in single FAB subtypes

Table 4  Frequencies of MLL duplication. Summary of reported studies