The t(6;9)(p23;q34)-DEK/CAN fusion occurs with an incidence of 1–5% in adult patients with acute myelogenous leukemia (AML) and tends to have an unfavorable prognosis at diagnosis. Due to the subtle appearance of this chromosome rearrangement, both initial detection and minimal residual disease (MRD) tracking by conventional karyotyping can be difficult. Unfortunately, no commercial or previously published fluorescence in situ hybridization (FISH) strategies exist for this recurrent anomaly. We have developed a highly sensitive assay using dual-color, double-fusion FISH (D-FISH), which can be used both for initial detection and MRD monitoring. We analyzed archived bone marrow samples from 15 patients with a previously identified t(6;9)(p23;q34) and 10 corresponding post-treatment samples. The results demonstrate that our D-FISH method effectively identified all abnormal samples, including a low-level MRD sample that was considered to be normal by conventional cytogenetic analysis. Normal value ranges were established from 30 negative controls to be <0.6% when 500 interphase nuclei were analyzed. The development of this sensitive D-FISH strategy for the detection of the t(6;9)(p23;q34) adds to the AML FISH testing repertoire, and is effective in the detection of low-level disease in post-treatment samples in these patients.
Recently, the World Health Organization (WHO) has linked recurrent chromosome abnormalities to specific subtypes of leukemia.1 Four recurrent chromosomal rearrangements associated with acute myelogenous leukemia (AML) have been described in detail, including t(8;21)(q22;q22), t(15;17)(q22;q21), inv(16)(p13q22), and t(11q23;var).1 Due to their relatively common prevalence and due to the diagnostic/prognostic significance of these abnormalities, commercial fluorescence in situ hybridization (FISH) probes have been available to test for these chromosomal rearrangements. However, although both the inv(3)(q21q26) and t(6;9)(p23;q34) are classically associated with AML, no commercially available dual-color, double-fusion FISH (D-FISH) probes are available to detect these anomalies.
Approximately 1–5% of patients with AML have a t(6;9)(p23;q34).2 The myeloid leukemia associated with the t(6;9)(p23;q34) has been most frequently classified by the French–American–British (FAB) system as AML-M2 and M4, though rarely M1 and RAEB.2, 3, 4, 5, 6 This translocation tends to occur in younger adults and is associated with an unfavorable prognosis at diagnosis. Although the t(6;9) is usually the sole cytogenetic aberration at diagnosis, additional karyotypic abnormalities are frequently identified during disease progression.2 The most common extra cytogenetic abnormalities observed are trisomies of chromosomes 8 and 13.
This reciprocal translocation involves recurrent breakpoints in the DEK gene on chromosome 6p23 and in the CAN (Cain, also known as NUP 214) gene on chromosome 9q34.3, 4 The recurrent breakpoints were localized to a 8 kb intronic region (icb-9) in the middle of the 140 kb CAN gene and a breakpoint localized to a 9 kb intronic region (icb-6) in the 40 kb DEK gene. The 3′ region of the CAN gene translocates to the 5′ region of the DEK gene on chromosome 6p23 via fusion of icb-9 to icb-6, and results in the transcription of a leukemia-specific, chimeric 5.5 kb mRNA transcript from the derivative chromosome 6.2, 4 The specific roles of the DEK and CAN proteins are poorly understood, but CAN appears to be involved in the transport of RNA and protein across the nuclear membrane, while DEK appears to be a nuclear protein which may be associated with chromatin structure.4, 7
Conventional cytogenetic analysis has been the gold standard for identification of the t(6;9)(p23;q34) in patients with AML. However, due to the relatively subtle appearance of this translocation (Figure 1), standard karyotypic analysis may overlook this abnormality, particularly at the lower resolution common to bone marrow preparations. Additionally, reverse transcription-polymerase chain reaction (RT-PCR) has been used successfully to detect this fusion.8, 9 With the advent of D-FISH as a cytogenetic testing modality, this technique has proven useful both in the initial detection of translocations and in monitoring of MRD. Patients with a reciprocal translocation typically generate a signal pattern of one red (1R), one green (1G), and two yellow fusion signals (2F). Generally, with the D-FISH strategy, the number of false-positive and false-negative cells approaches zero.10
Since the t(6;9)(p23;q34) occurs as an apparently balanced reciprocal translocation, the opportunity to utilize a D-FISH strategy seems appropriate in detecting and monitoring this subtle translocation. In an effort to expand our FISH panel testing for AML, we developed home-brew FISH probes for the DEK and CAN genes to evaluate a novel D-FISH approach to identify the t(6;9)(p23;q34) in neoplastic cells from patients with AML.
Materials and methods
Following the Internal Review Board approval, we conducted a retrospective review of our Cytogenetics database at Mayo Clinic to identify bone marrow samples in which a t(6;9)(p23;q34) was identified by conventional cytogenetic techniques.11 We selected 15 patient samples with the t(6;9)(p23;q34) and 10 corresponding post-treatment samples. In addition 30 normal bone marrow specimens were included as negative controls to establish normal cutoffs for our FISH probe strategy. The 25 patient specimens and 30 normal specimens were blinded for the study (55 total specimens).
Direct-labeled FISH probes were designed from bacterial artificial chromosomes (BACs) and validated according to standard methods.12 For the DEK gene, the two clones utilized were RP11-204B7 (147 kb) and RP1-298J15 (109 kb) (Figure 2a). For the CAN gene, the four clones utilized were RP11-143H20 (50 kb), RP11-544A12 (197 kb), RP11-643E14 (112 kb), and RP11-334J6 (147 kb) (Figure 2b). Figure 2c and d illustrates the translocated probe regions for both DEK and CAN on the derivative chromosomes 6 and 9 yielding yellow fusion signals at both breakpoints. The DEK and CAN BACs were labeled with Spectrum Orange-dUTP and Spectrum Green-dUTP, respectively. The CAN probe was intentionally made larger since this probe was labeled in SpectrumGreen™ (Vysis Inc., Downers Grove, IL, USA), which characteristically produces a dimmer signal.
Bone marrow cell pellets were stored at −70°C in methanol:glacial acetic acid (2:1) fixative. After a change of fixative, the fixed-cell pellet suspensions were manually dropped onto microscope slides and checked by phase contrast microscopy to verify appropriate cellularity. Samples were then subjected to standard FISH pretreatment, hybridization, and fluorescence microscopy methods.13
FISH signal patterns
Figure 3 consists of interphase nuclei with representative signal patterns. The DEK probe on chromosome 6p23 is labeled in red (R) and the CAN probe on chromosome 9q34 is labeled in green (G). Normal interphase nuclei have a 2R2G signal pattern (Figure 3a). In some normal interphase nuclei, a single fusion (F) signal may be observed due to coincidental overlap of R and G signals. This yields a signal pattern of 1R1G1F (Figure 3b). For the majority of patients with a t(6;9)(q23;q34), abnormal interphase nuclei will display a D-FISH signal pattern of 1R1G2F (Figure 3c). For some patients, variable signal patterns of 2R2G1F (Figure 3d), 1R1G1F (Figure 3b), 1R2G1F (Figure 3e), and 2R1G1F (Figure 3f) may be observed due to complex translocations, atypical breakpoints, and/or loss of DNA at the breakpoint junctions.
Results and discussion
FISH probe validation
In this blinded study, we analyzed 30 normal bone marrow specimens as negative controls and 25 specimens from 15 patients with the t(6;9)(p23;q34) (Table 1 , negative control data not shown). We expected the D-FISH signal pattern of 1R1G2F to predominate in all of the abnormal specimens, as each patient had an apparently ‘balanced’ t(6;9)(p23;q34) by conventional chromosome analysis. In addition, we postulated that MRD in post-treatment samples with normal chromosome results could possibly be detected by FISH due to its increased sensitivity. The home-brew probes successfully hybridized to all specimens and produced bright signals for both the DEK and CAN gene regions. After the study was unblinded and compared to the chromosome results, the D-FISH method correctly identified all 15 patient diagnostic specimens and all 30 negative control specimens, yielding a clinical sensitivity and specificity of 100%.
Table 1 summarizes the results for the diagnostic and post-treatment samples for the 15 patients. The abnormal reference range for the abnormal diagnostic samples was 85.6–99.6%. All 15 diagnostic samples displayed a signal pattern of 1R1G2F, with the exception of sample 6, which had a signal pattern of 1R1G1F. For the 10 corresponding post-treatment samples from patients 11–15, FISH deemed eight of the 10 serial samples abnormal with a signal pattern of 1R1G2F, with the exception of sample 13c (signal pattern 1R2G1F). These eight samples were considered to be concordant with chromosome results, with the exception of sample 13b, which was cytogenetically normal. For the two remaining serial samples 12b and 15d, FISH results were within normal limits and concordant with the chromosome results.
At the conclusion of the study, the results of the 30 negative controls were used to calculate the normal cutoff ranges for nuclei displaying false-positive signal patterns. For each signal pattern, we used the maximum number of false-positive nuclei in 500 cells from a single negative control to determine the calculated normal cutoffs (Table 2 ). Since a small number of negative controls were analyzed, we used a one-sided, binomial distribution with a 95% confidence interval. This method provides the upper bound (or percentage) of false-positive nuclei for an individual subject, rather than the mean of many subjects (for further explanation, see Fan (2002)).14 In the current study, the signal patterns of 1R1G2F, 2R2G1F, 1R2G1F, and 2R1G1F had zero false-positive nuclei observed. These results indicate that the upper bound normal cutoff is reached when three cells or greater (>0.6%) are identified. For the single fusion pattern of 1R1G1F, one of the negative controls had 35 false-positive cells identified in the 500 cells scored. Using Table 2, this result indicates that the upper bound normal cutoff is reached when 47 cells or greater (>9.4%) are observed. The significantly higher normal cutoff for 1R1G1F is due to random overlap in normal nuclei. The implications of a low analytical sensitivity for the 1R1G1F signal pattern will be a diminished ability to detect MRD for post-treatment samples of patients that harbor this uncommon signal pattern. Although analysis of 500 nuclei will detect a high proportion of cases with MRD, our laboratory manually analyzes up to 6000 nuclei for cases with a history of a specific hematologic translocation. Based on the cutoff determination method described above, analysis of 6000 nuclei will improve the analytical sensitivity from 0.6% disease detection for 500 nuclei to 0.05% (three cells in 6000) disease detection.
Atypical FISH signal patterns
Complex translocations, atypical breakpoints, and DNA loss at breakpoint junctions contribute to the variety of signal patterns observed in interphase nuclei. The clinical significance of DNA loss at breakpoints which generate these atypical signal patterns is unclear. The diagnostic specimen for patient 6 displayed only one fusion signal, yielding a pattern of 1R1G1F in 94.6% of 500 nuclei. In two previously published reports, the derivative chromosome 9 was lost in a subset of patients which would produce this pattern.2, 3 In addition, we would not expect the derivative chromosome 6 to be lost, since it was previously found to drive leukemogenesis.3, 4 However, retrospective review of the karyotype for patient 6 confirmed the presence of a ‘balanced’ reciprocal t(6;9)(p23;q34). After a review of metaphases with FISH, we discerned that the residual fusion was present on the derivative chromosome 6 (Figure 4a). The observations of the FISH fusion signal on the derivative chromosome 6 and of the karyotypic finding of an apparently balanced t(6;9)(p23;q34) suggest that patient 6 has a cryptic loss of the translocated DEK and residual CAN DNA on the derivative chromosome 9.
Diagnostic sample 13a had an apparently ‘balanced’ reciprocal t(6;9)(p23;q34) by chromosome analysis and an expected signal pattern of 1R1G2F. Serial sample 13b was normal by conventional chromosome studies but abnormal by FISH (discussed under MRD section). Chromosome results for serial sample 13c indicate a relapse of the t(6;9)(p23;q34) along with additional chromosome abnormalities consisting of a del(5q) and an add(7q), which represent clonal progression of this patient's myeloid leukemia. Interestingly, sample 13c displayed the result of an atypical signal pattern of 1R2G1F in 78.8% of 500 interphase nuclei. Retrospective review of the karyotypes for samples 13a and 13c suggest that the observed t(6;9)(p23;q34) in both samples is cytogenetically identical. Unfortunately, we were unable to do metaphase FISH analysis to determine the nature of the atypical FISH pattern due to low sample volume. Our results suggest deletion of DEK sequences either from the derivative chromosome 6 or the derivative chromosome 9. Since the derivative chromosome 6 has been reported to drive leukemogenesis, we assume the translocated DEK DNA on the derivative chromosome 9 was subsequently deleted in this sample. The clinical significance of the initial occurrence of a balanced translocation with subsequent loss of junctional DNA sequences is unknown. This observation suggests that the sequences involved in this translocation were unstable and prone to loss.
The amenability of FISH is its usefulness in identification of variant translocations and/or breakpoints not otherwise elucidated by conventional cytogenetics. This ability has been well documented with the t(9;22)(q34;q11.2) in CML.15, 16 Similarly, in AML, atypical FISH signal patterns indicating sequence deletions have been reported at the breakpoint junctions in patients with t(8;21)(q22;q22), inv(16)(p13q22), and t(11q23;var).17, 18 Identification of these alternate signal patterns is important, because these patients may respond differently to treatment. Although patients with DEK/CAN fusion in AML are rare, the observation of atypical FISH signal patterns in additional patients may discern clinical significance between patients with various signal patterns.
Minimal residual disease (MRD)
The ability to detect and quantify MRD in post-treatment samples from patients with hematologic disorders is necessary to monitor response to treatment. Both RT-PCR and FISH have been developed to monitor MRD for the most common translocations in hematologic disorders, and have shown similar efficacy based on the laboratory's proficiency with these techniques. Our laboratory has become proficient in the application of FISH methods for MRD detection, as we routinely analyze up to 6000 nuclei for FISH tests utilizing the D-FISH strategy. These results can be used to compare post-treatment samples for response to therapy. This approach has been used in the clinical setting, thus proving its reliability as a molecular cytogenetic method.10, 19
In this study, we included three post-treatment samples (12b, 13b, and 15d) with normal chromosome results to determine if MRD could be identified by FISH. Samples 12b and 15d had zero cells (0%) and two cells (0.03%), respectively, identified out of 6000 nuclei with DEK/CAN fusion. Although these results are concordant with chromosome results, we postulate that sample 15d was in very early cytogenetic relapse with very low levels of disease since the D-FISH test theoretically produces no false-positive signal patterns of 1R1G2F. For serial sample 13b, FISH identified 1.8% of 500 nuclei with DEK/CAN fusion. The finding of MRD for this patient may have predicted the need for additional therapeutic intervention prior to this patient's subsequent relapse.
Identification of a t(6;9)(p23;q34) and corresponding DEK/CAN gene fusion is associated with an unfavorable prognosis in patients with AML. Effective detection methods are essential to recognize this subtle translocation. D-FISH is a powerful methodology that allows both the initial detection and MRD tracking for translocations. The home-brew D-FISH probes we designed to detect DEK/CAN gene fusion performed very well in both diagnostic and MRD specimens, allowing disease detection down to >0.05% if 6000 nuclei are analyzed. In closing, we plan to incorporate the t(6;9)(p23;q34) D-FISH test into our AML FISH panel for application in patients with a referral of AML that have putatively normal or unsuccessful chromosome results and in patients for MRD detection. The addition of this FISH test further adds to the list of classic AML-associated translocations for which both diagnostic and MRD testing can be achieved by molecular cytogenetic methods.
Jaffe ES, Harris NL, Stein H, Vardiman JW . World Health Organization Classification of Tumours: Pathology & Genetics of Tumours of Haematopoietic and Lymphoid Tissues. Lyon, France: International Agency for Research on Cancer, 2001, pp 75–107.
Soekarman D, von Lindern M, Daenen S, de Jong B, Fonatsch C, Heinze B et al. The translocation (6;9)(p23;q34) shows consistent rearrangement of two genes and defines a myeloproliferative disorder with specific clinical features. Blood 1992; 79: 2990–2997.
von Lindern M, Poustka A, Lerach H, Grosveld G . The (6;9) chromosome translocation, associated with a specific subtype of acute nonlymphocytic leukemia, leads to aberrant transcription of a target gene on 9q34. Mol Cell Biol 1990; 10: 4016–4026.
von Lindern M, Fornerod M, van Baal S, Jaegle M, de Wit T, Buijs A et al. The translocation (6;9), associated with a specific subtype of acute myeloid leukemia, results in the fusion of two genes, dek and can, and the expression of a chimeric, leukemia-specific dek–can mRNA. Mol Cell Biol 1992; 12: 1687–1697.
Heim S, Kristoffersson U, Mandahl N, Mitelman F, Bekassy AN, Garwicz S et al. High resolution banding analysis of the reciprocal translocation t(6;9) in acute nonlymphocytic leukemia. Cancer Genet Cytogenet 1986; 22: 195–201.
Soekarman D, von Lindern M, van der Plas DC, Selleri L, Bartram CRI, Martiat P et al. Dek–Can rearrangement in translocation (6;9)(p23;q34). Leukemia 1992; 6: 489–494.
Fornerod M, Boer J, van Baal S, Morreau H, Grosveld G . Interaction of cellular proteins with the leukemia specific fusion proteins DEK–CAN and SET–CAN and their normal counterpart, the nucleoporin CAN. Oncogene 1996; 13: 1801–1808.
Hamaguchi H, Nagata K, Yamamoto K, Fujikawa I, Kobayashi M, Eguchi M . Establishment of a novel human myeloid leukaemia cell line (FKH-1) with t(6;9)(p23;q34) and the expression of dek–can chimaeric transcript. Br J. Haematol 1998; 102: 1249–1256.
Boer J, Mahmoud H, Raimondi S, Grosveld G, Krance R . Loss of the DEK–CAN fusion transcript in a child with t(6;9) acute myeloid leukemia following chemotherapy and allogeneic bone marrow transplantation. Leukemia 1997; 11: 299–300.
Dewald GW, Wyatt WA, Juneau AL, Carlson RO, Zinsmeister AR, Jalal SM et al. Highly sensitive fluorescence in situ hybridization method to detect double BCR/ABL fusion and monitor response to therapy in chronic myeloid leukemia. Blood 1998; 91: 3357–3365.
Spurbeck JL, Carlson RO, Allen JE, Dewald GW . Culturing and robotic harvesting of bone marrow, lymph nodes, peripheral blood, fibroblasts, and solid tumors with in situ techniques. Cancer Genet Cytogenet 1988; 32: 59–66.
Pardanani A, Ketterling RP, Brockman SR, Flynn HC, Paternoster SF, Shearer BM et al. CHIC2 deletion, a surrogate for FIP1L1–PDGFRA fusion, occurs in systemic mastocytosis associated with eosinophilia and predicts response to imatinib mesylate therapy. Blood 2003; 102: 3093–3096.
Brockman SR, Paternoster SF, Ketterling RP, Dewald GW . New highly sensitive fluorescence in situ hybridization method to detect PML/RARA fusion in acute promyelocytic leukemia. Cancer Genet Cytogenet 2003; 145: 144–151.
Fan YS (ed). Methods in Molecular Biology. Totowa, NJ: Humana Press Inc., 2002; 204: 334–338.
Smoley SA, Brockman SR, Paternoster SF, Meyer RG, Dewald GW . A novel tricolor, dual-fusion fluorescence in situ hybridization method to detect BCR/ABL fusion in cells with t(9;22)(q34;q11.2) associated with deletion of DNA on the derivative chromosome 9 in chronic myelocytic leukemia. Cancer Genet Cytogenet 2004; 148: 1–6.
Sinclair PB, Nacheva EP, Leversha M, Telford N, Chang J, Reid A et al. Large deletions at the t(9;22) breakpoint are common and may identify a poor-prognosis subgroup of patients with chronic myeloid leukemia. Blood 2000; 95: 738–744.
Kolomietz E, Al-Maghrabi J, Brennan S, Karaskova J, Minkin S, Lipton J et al. Primary chromosomal rearrangements of leukemia are frequently accompanied by extensive submicroscopic deletions and may lead to altered prognosis. Blood 2001; 97: 3581–3588.
Godon C, Proffitt J, Dastugue N, Lafage-Pochitaloff M, Mozziconacci M-J, Talmant P et al. Large deletions 5′ to the ETO breakpoint are recurrent events in patients with t(8;21) acute myeloid leukemia. Leukemia 2002; 16: 1752–1754.
Wang YL, Bagg A, Pear W, Nowell PC, Hess JL . Chronic myelogenous leukemia: laboratory diagnosis and monitoring. Genes Chromos Cancer 2001; 32: 97–111.
About this article
Cite this article
Shearer, B., Knudson, R., Flynn, H. et al. Development of a D-FISH method to detect DEK/CAN fusion resulting from t(6;9)(p23;q34) in patients with acute myelogenous leukemia. Leukemia 19, 126–131 (2005). https://doi.org/10.1038/sj.leu.2403557
- acute myelogenous leukemia