1. ¹Leukaemia Research Fund Cytogenetics Group, Cancer Sciences Division, University of Southampton, Southampton, UK
2. ²Institute of Comparative Medicine, Glasgow University Veterinary School, Glasgow, UK
3. ³The Orchid Cancer Appeal, Cancer Research UK Medical Oncology Unit, London, UK
4. ⁴Structural Cell Biology, Paterson Institute for Cancer Research, Christie Hospital (NHS) Trust, Manchester, UK

Correspondence: Dr CJ Harrison, Leukaemia Research Fund Cytogenetics Group, Cancer Sciences Division, University of Southampton, MP 822 Duthie Building, Southampton General Hospital, Southampton SO16 6YD, UK. Fax: +44 23 8079 6432; E-mail: harrison@soton.ac.uk

Received 5 June 2003; Accepted 18 February 2004; Published online 1 April 2004.

TO THE EDITOR

The Leukaemia Research Fund UK Cancer Cytogenetics Group Karyotype Database in Acute Leukaemia (Database)¹ would like to inform readers of a new finding. We observed amplification involving the ABL gene in acute lymphoblastic leukemia (ALL), which appeared as multiple signals by fluorescence in situ hybridization (FISH) in eight patients with T-lineage ALL. Gene amplification is a frequent finding in a wide range of tumors but has been rarely described in acute leukemia. In cytogenetic preparations, it is usually seen intrachromosomally as homogeneously staining regions (HSR) or extrachromosomally as double minutes (dmin). FISH studies have revealed the genes involved in the amplified regions. Amplifications of the MYC and MLL genes have been previously reported in acute myeloid leukemia (AML),^2,3,4 and of AML1 and MLL in ALL.^3,5 Although amplification of the BCR/ABL fusion gene has been described in cases of chronic myeloid leukemia (CML) treated with imatinib mesylate, seen both as HSRs⁶ and dmins,⁷ amplification of ABL alone is rare. A six-fold and a 15-fold amplification of the gene have been described in the CML-derived cell line, K562,⁸ and in a patient with CML in lymphoid blast crisis, respectively.⁹ The only other FISH report of amplification involving ABL was in three cases of secondary AML.¹⁰

Chromosomal analysis of diagnostic bone marrow samples in seven patients and the diagnostic and relapse samples in one patient (4711) was carried out in the UK regional cytogenetics laboratories and reviewed by the Database cytogeneticists. Karyotypes were described according to the International System of Human Cytogenetic Nomenclature.¹¹ Metaphase and interphase FISH analysis was carried out on the same nine fixed-cell suspensions as used for cytogenetic analysis. The dual color probe kits, LSI^® BCR/ABL, LSI^® TEL/AML1 ES and LSI^® MLL (Vysis, UK), were hybridized to cells from the diagnostic samples of the eight patients as part of a routine screen for the BCR/ABL and TEL/AML1 fusions and rearrangements of the MLL gene, respectively. The whole chromosome painting probe for chromosome 9 (wcp9) (STAR*FISH, Cambio, UK) was sequentially hybridized onto metaphases from three patients (4273, 5980 and 6099) following the BCR/ABL probe. Selected centromere probes were applied to interphase cells from patients with a failed or normal cytogenetic result (3183, 3940 and 5980) to exclude hidden numerical chromosomal changes. Multiplex FISH (M-FISH), using the SpectraVysion 24 color painting kit (Vysis, UK), was carried out on metaphases from four patients (3550, 4273, 6099 and 6905) to define their complex karyotypes. In order to confirm that the amplified region included the ABL gene, three PAC probes mapping 5' to 3' from intron 1 (dJ1132H12, dJ913J14 and dJ888H1, Dr Rocchi, Bari, Italy) and two mapping to downstream exons (including exon 11) (dJ835J22 and dJ1146L15, Dr Rocchi, Bari, Italy) were hybridized individually onto interphase and metaphase cells from the nine samples, as previously described.⁵ In patient 4273, the amplified ABL signals were counted automatically using a Cytovision SPOT AX system (Applied Imaging International Ltd, UK).

Southern blotting and hybridization analysis was carried out on genomic DNA from the diagnostic bone marrow of patient 3183, as previously described.¹² DNA from the cell line, K562, two other childhood ALL patient samples (2718, 3013) and normal leukocytes were used as controls. Gene-specific genomic microarray (array CGH) analysis was performed on the same genomic DNA sample from patient 3183, using an Array300 chip (Abbot Diagnostics, UK) and the GenoSensor Reader System (Abbot Diagnostics, UK) as previously reported.¹³ This array allows identification of copy number changes to 287 genes commonly altered in human tumors.

Among 280 patients with T ALL entered to the UK Medical Research Council ALL treatment trials, ALL97 (1997–2002, for children aged 1–18 years) and UKALLXII (1992–present, for adults aged 15–55 years), a total of eight patients, five children and three adults, showed multiple signals for the ABL gene with the BCR/ABL probe kit (Figure 1a). Amplified signals were present in both the diagnostic and relapse samples from one patient (4711). The estimated incidence of this ABL amplification among childhood T ALL cases appeared to be lower at 5/210 (2.3%) compared with 3/70 (4.3%) adults with T ALL. This phenomenon was not observed among more than 1600 childhood and 300 adult patients with B-lineage ALL screened with the same probe set.

Figure 1.

(a) Representative interphases and a tetraploid metaphase from patient 4273 hybridized with LSI^® BCR/ABL probe. BCR signals (green spots arrowed) indicate the four copies located to BCR loci on chromosomes 22. Multiple red signals indicate amplification of ABL. The additional signals are randomly distributed over the metaphase and are not specifically hybridized to the chromosomes. Sequential hybridization with wcp9 and superimposing of images indicated four copies of chromosome 9 (green) with the ABL signals (red) in the expected location on 9q in addition to the extra signals. (b) Cells from patient 3940 hybridized with the PAC dJ114L15 indicating amplification involving ABL exon 11. (c) The same interphase cell from patient 4273 hybridized with the LSI^® BCR/ABL probe: (i) shows the original image as captured by the SPOT AX system and (ii) demonstrates how SPOT AX can identify and enumerate the signals from the BCR and ABL probes separately.

Full figure and legend (169K)

Clinical and cytogenetic details of these eight patients are given in Table 1. Five were children aged 1–15 years (median 12 years), while three were adults aged 18–42 years (median 36 years). Apart from the one infant, this abnormality was associated with an older age group. In addition to the T ALL immunophenotype, all patients had L1/L2 morphology. Event-free survival was generally poor. Three of the five children died within 14 months of diagnosis, two following early bone marrow relapse and one in remission. One adult patient relapsed at 21 months. There was one child survivor at 34 months after diagnosis. Follow-up for the remaining patients was short.

Table 1 - Clinical, survival and cytogenetic details of the eight T ALL patients with amplification of ABL.

Full table

Cytogenetic analysis was successful in six of the eight patients with additional ABL signals. In five of them the karyotype was abnormal. The only established chromosomal change was t(10;14)(q24;q11) in two adult patients (4711, 6905). In those patients with metaphases, sequential painting with wcp9 revealed copies of ABL in their expected positions on chromosomes 9 in addition to the extra signals as well as the expected number of BCR signals (Figure 1a). There was no evidence of the BCR/ABL fusion in any patient, while two signals for each of the genes were seen in the normal cells. TEL/AML1 and MLL screening produced normal results.

Five specific PAC probes confirmed that the amplified regions contained at least part of the ABL gene (Figure 1b), although there was variation between patients. Two patients (3183 and 3550) showed identical amplification of all five probes. Three patients (4711, 6099 and 6905) had amplified signals for four probes; dJ1132H12 was not amplified and showed the normal two signals. An identical pattern to that of diagnosis was observed in the relapse sample from patient 4711 taken 21 months later. The remaining three patients (3940, 4273 and 5980) only showed multiple copies of the probes dJ888H11, dJ835J22 and dJ1146L15. This demonstrated that the amplified regions of intron 1 but not of exon 11 were different between patients.

ABL amplification was present in the majority of interphase cells examined in all but one patient (6905) (Table 1). The number of additional ABL signals varied from cell to cell in all eight patients. By eye they were difficult to enumerate accurately at both metaphase and interphase. This was resolved in one patient (4273) by automated signal counting across several focal planes, which revealed a range of 8–57 (median 18) signals (Figure 1c).

Prepared slides from five patients were pretreated with RNAse (0.1 mg/ml) (Sigma, UK) prior to hybridization with the PAC probes to ensure that all residual traces of RNA had been removed from the cells. In one patient, increasing concentrations (0.05–0.4 mg/ml RNAse) were used. Identical signal patterns were observed for treated and untreated samples, which ruled out that the probes were detecting mRNA.

High molecular weight genomic DNA was available for molecular analysis in one patient only (3183). Southern analysis revealed a germline 24.8 kb ABL-specific band (Figure 2). The intensity of this band was significantly increased in patient 3183 and the K562 cell line. K562 is known to have an approximately six-fold amplification of ABL.⁸ In our patient sample (3183), ABL was amplified at least five-fold in comparison with K562, taking into consideration that only 67% of the cells within the bone marrow sample showed evidence of amplification. Examination of the ethidium bromide gel confirmed that the increased signal intensity in K562 and patient 3183 was not a loading artefact. An additional nonspecific 9.8 kb band was also detected. The intensity of this band was constant across the samples analyzed and served as a DNA loading control. The other two patient samples (2718 and 3013) had two and three copies of the ABL signal, respectively, as detected by FISH. There was no discernable difference in the germline band intensity between them and the leukocyte control.

Figure 2.

Southern blot analysis of genomic DNA from patient 3183, positive and negative controls using a probe against ABL. Controls were commercially purchased normal leukocyte DNA, two samples from childhood ALL patients (2718, 3013) and the K562 cell line. GL: germline band; NS: band from nonspecific fragment. This band fragment shows no difference in intensity across the samples.

Full figure and legend (64K)

The increased ABL copy number demonstrated by FISH in patient 3183 was also supported by a high-level gain detected with the array CGH (Table 2). This technique has previously been shown to provide accurate analysis of gene copy number alteration in human tumors.^14,15 The value of 2.63 (P<0.001) for the ABL gene was significantly higher than copy number gains for other genes represented in this array. Thus, these data confirm that the high number of FISH signals observed represents amplification of the ABL gene. They also demonstrate an accurate correlation between the FISH and CGH array result at several other gene loci; BCR, MLL, TEL and AML1 (Table 2).

Table 2 - FISH and array CGH of patient 3183.

Full table

The additional ABL signals detected by FISH did not appear to be hybridized to any particular chromosome, but were randomly distributed throughout the cells. Metaphases prepared directly onto slides without any subsequent hybridization showed no visible evidence of dmins when stained with Giemsa (Sigma, UK) or Propidium Iodide (Vysis, UK). Scanning electron microscopy (SEM), using a previously described procedure,¹⁶ which had successfully revealed dmins of submicroscopic size in a tumor cell line, did not demonstrate their presence in the two patients investigated (4273 and 3550).

Amplification involving ABL was discovered by chance during routine FISH screening of ALL patients for the presence of the BCR/ABL fusion. Sequence-specific probes confirmed that the amplified region contained at least part of the ABL gene. It was found exclusively in T-lineage ALL and, apart from one infant, occurred in patients >10 years old. Although suitable genomic DNA was available for only one patient, this provided convincing supporting molecular evidence of amplification by Southern analysis and array CGH. Metaphase FISH analysis revealed that the amplification was extrachromosomal in nature, although there was no visible evidence of dmins. Amplified c-MYC sequences have been found on submicroscopic circular extrachromosomal DNAs called episomes in the HL-60 cell line¹⁷ and amplified DHFR genes have been observed on homogeneous extrachromosomal DNA molecules of approximately 650 kb in mammalian cells.¹⁸ Thus, the location of the amplified ABL sequences to comparable submicroscopic extrachromosomal structures cannot be ruled out. Whether ABL amplification is a manifestation of a fundamentally important biological process remains to be proved. However, the recurrent association with T ALL that we have demonstrated is a significant new observation.