The RS4;11 cell line as a model for leukaemia with t(4;11)(q21;q23): Revised characterisation of cytogenetic features

Abstract Background Haematological malignancies harbouring rearrangements of the KMT2A gene represent a unique subtype of leukaemia, with biphenotypic clinical manifestations, a rapid and aggressive onset, and a generally poor prognosis. Chromosomal translocations involving KMT2A often cause the formation of oncogenic fusion genes, such as the most common translocation t(4;11)(q21;q23) producing the KMT2A‐AFF1 chimera. Aim The aim of this study was to confirm and review the cytogenetic and molecular features of the KMT2A‐rearranged RS4;11 cell line and put those in context with other reports of cell lines also harbouring a t(4;11) rearrangement. Methods and Results The main chromosomal rearrangements t(4;11)(q21;q23) and i(7q), described when the cell line was first established, were confirmed by fluorescence in situ hybridisation (FISH) and 24‐colour karyotyping by M‐FISH. Additional cytogenetic abnormalities were investigated by further FISH experiments, including the presence of trisomy 18 as a clonal abnormality and the discovery of one chromosome 8 being an i(8q), which indicates a duplication of the oncogene MYC. A homozygous deletion of 9p21 containing the tumour‐suppressor genes CDKN2A and CDKN2B was also revealed by FISH. The production of the fusion transcript KMT2A‐AFF1 arising from the der(11)t(4;11) was confirmed by RT‐PCR, but sequencing of the amplified fragment revealed the presence of multiple isoforms. Two transcript variants, resulting from alternative splicing, were identified differing in one glutamine residue in the translated protein. Conclusion As karyotype evolution is a common issue in cell lines, we highlight the need to monitor cell lines in order to re‐confirm their characteristics over time. We also reviewed the literature to provide a comparison of key features of several cell lines harbouring a t(4;11). This would guide scientists in selecting the most suitable research model for this particular type of KMT2A‐leukaemia.


| INTRODUCTION
Leukaemia harbouring rearrangements of the KMT2A gene (formerly known as MLL/mixed-lineage leukaemia and also known as HRX or TRX1) represent a unique subtype of acute leukaemia, characterised by a rapid and aggressive onset with generally poor prognosis. KMT2A rearrangements can give rise to different cellular phenotypes, with affected cells showing an interesting lineage heterogeneity, hence the designation of "mixed-lineage". Rearrangements involving KMT2A are often found in de novo and DNA topoisomerase II inhibitor therapy-related myeloid and lymphoblastic acute leukaemias, with varying incidence according to type and age. 1 Overall, KMT2A rearrangements account for 10% of all acute leukaemias 2 but are predominantly found in infants between the age of 0 and 2 diagnosed with acute lymphoblastic leukaemia (ALL; 70%-80% of cases 3,4 ) and in therapy-related acute myeloid leukaemia (AML) patients (up to 70% of cases). 5 Chromosomal translocations are common rearrangements in KMT2A-leukaemia, 6 whereby the exchange of genetic material between chromosomes brings the N-terminus of KMT2A to fuse inframe with the C-terminus of a partner gene. 7,8 The breakpoint region of KMT2A covers approximately 8 kbp between exons 7 and 11. 9 More than 90 partner genes for KMT2A have been identified, forming the so-called "MLL recombinome." The most common translocation partners are AFF1 (previously known as AF4) on 4q21, MLLT3 (AF9) on 9q22, ELL on 19p13.1, MLLT1 (ENL) on 19p13.3, MLLT10 (AF10) on 10p12, and MLLT4 (AF6) on 6q27, giving rise to t(4;11)(q21;q23), t(9;11)(q22;q23), t(11;19)(q23;p13.1), t(11;19)(q23;p13.3), t(10;11) (p12;q23), and t(6;11)(q27;q23), respectively. 1 The in-frame fusion of KMT2A and a partner gene produces chimeric proteins with oncogenic activity, which largely depends on the retained domains of KMT2A and the characteristics of the fusion partner. 10 The most common translocation in the MLL recombinome is the t(4;11)(q21;q23), which produces the KMT2A-AFF1 chimeric protein. 1,11 The phenotype is mainly B-ALL, with rare cases of AML.
These leukaemic cells possess a biphenotypic profile, as they coexpress lymphoid and myeloid markers while maintaining a lymphoblastic morphology, suggesting that the original malignant clone arises from an early lymphoid/myeloid precursor. 12 Leukaemia initiation by KMT2A rearrangements is thought to occur via an improper expression and regulation of Hox genes. 13 The KMT2A protein is a homologue of the trithorax protein in Drosophila melanogaster 14 and functions as a transcriptional activator and regulator of Hox genes during embryogenesis and haematopoiesis. 15,16 AFF1 is a nuclear protein acting as transcriptional regulator involved in haematopoietic development of lymphoid precursors. 17 The KMT2A-AFF1 fusions are capable of initiating and maintaining an erroneous programme of transcription with oncogenic consequences. 18 It is a topic of debate whether KMT2A fusions alone are sufficiently powerful to cause the disease, contradicting the "multi-hit model" that applies to most leukaemias. 19 Supporting the prenatal origin of leukaemia, alterations of the KMT2A gene have been documented in utero. 20 One of the most remarkable features of infant KMT2A-leukaemia is the extraordinarily short latency and the limited number of secondary somatic mutations, 21 indicating that only a small number of additional events may be necessary to initiate the malignancy, if at all. 22,23 Mutations in FLT3, KRAS, and NRAS have been proposed as "second hits" drivers for leukaemogenesis of KMT2Aleukaemias. 24,25 A reliable in vivo model faithfully mimicking the disease is still lacking (reviewed in Ottersbach et al 26 ). Interestingly, murine models so far have achieved a transient production of KMT2A-AFF1 proteins but failed to express the phenotype observed in humans. [27][28][29] Only recently, the development of such models is becoming increasingly closer to clinical phenotypes. 30,31 Therefore, in vitro models have been at the forefront of research into KMT2A-leukaemia. Although at least 16 cell lines with the t(4;11) have been described, only four have been appropriately authenticated. 32,33 In this article, we focus on the revisitation of the RS4;11 cell line, which was first established by Stong et al 34 from a 32-year-old female patient with relapsed ALL. The initial karyotype described a t(4;11)(q21;q23) and the presence of an i(7q).
Our work focuses on the characterisation of cytogenetic and molecular aspects of RS4;11 in comparison with previous work from other groups. As karyotype evolution is common in extended cell cultures, 32  penicillin/streptomycin (100 UmL −1 /μgmL −1 ) (Gibco), and incubated at 37°C in 5% CO 2 . Cells were passaged every 48 hours.

| RT-PCR and cloning of the KMT2A-AFF1 fusion
Total RNA was extracted from cell cultures using TRIzol Reagent and converted into cDNA using random hexamers and SuperScript III reverse transcriptase, according to manufacturer's instructions
The involvement of the KMT2A gene was further confirmed by FISH using a dual-colour break-apart probe encompassing the KMT2A locus at 11q23.3. One normal KMT2A allele could be seen as a yellow fusion signal, whereas disruption of the KMT2A region would result in one red and one green signal present on the der(11) and der(4), respectively ( Figure 3).

| Confirmation of an isochromosome 7q by FISH
The presence of an i(7q) was detected by FISH using arm-specific chromosome paints for 7q (red) and 7p (green), and single locus probes for bands 7q22 (red) and 7q36 (green). Arm-specific probes showed the complete coverage of one chromosome 7 in red, corresponding to the long arm ( Figure 2C). For single locus probes, this was confirmed by the presence of two additional signals for 7q22 and 7q36 on one chromosome 7 ( Figure 2D).

| Identification of an isochromosome 8q leading to duplication of 8q24
M-FISH ( Figure 1) and FISH using WCP8 ( Figure 2E) revealed two copies of chromosome 8 different in size and centromere position.
FISH using a single locus DNA probe for 8q24.3 (RP11-195E4) was carried out to investigate a possible duplication of this locus. Signals specific for this region were visible on opposite sides of the centromere, suggesting the presence of an isochromosome 8q ( Figure 2F).

| Homozygous deletion of 9p21 detected by FISH
The dual-colour probe XL CDKN2A/2B highlighted the presence of both chromosome 9 centromeres. However, lack of both signals for 9p21 revealed a homozygous deletion of that region ( Figure 2G,H).
Additional numerical abnormalities have been previously described in RS4;11, notably trisomy 8 and 18, and monosomy X. 42,43 In our study, although FISH using whole chromosome paints for chromosomes 8, 18, and X did not confirm these aneuploidies, two separate clones with +8 and +18, respectively, were identified by M-FISH.
Trisomy of chromosomes 8 and 18 can arise in extended cell cultures and seems to confer a proliferative advantage by an increased gene dosage effect. 44,45 Trisomy 8 is present in the KMT2Arearranged cell line MV-4-1 46 and in a number of myeloid leukaemia cell lines such as K-562, 47 SKK-1, 48 and GDM-1. 49 Trisomy 18 is more common in cell lines derived from solid tumours. 44 In leukaemia patients, trisomy 8 is usually described in therapy-related leukaemia or as a secondary clonal event, 50,51 with a possible role in disease progression rather than in primary leukaemogenesis. 52 Although rare, trisomy 18 in leukaemia is mainly found in conjunction with other abnormalities, such as trisomy 12 or 16 in chronic lymphocytic leukaemia (CLL), and is also regarded as an event of clonal evolution. 53,54 We report the consistent presence of an i(8q) in RS4;11, which has not been previously described. The formation of the i(8q) results in a duplication of the proto-oncogene c-Myc mapping at 8q24.2, which has been shown to provide selective growth advantage in vitro. [55][56][57] In AML and ALL, i(8q) is considered to arise as a secondary clonal abnormality contributing to disease progression and is often found in conjunction with complex karyotypes. 58,59 We confirmed the homozygous deletion at 9p21, a locus containing The del(9p21) in conjunction with t(4;11) or other KMT2A translocations are seen but at a lower frequency than in other cytogenetic subgroups such as t(1;19) and t(9;22)/BCR-ABL, indicating that the inactivation of CDKN2A/2B is not indispensable for the malignant phenotype. 64 At the molecular level, known breakpoints on KMT2A span the region between exon 7 and exon 11, and in AFF1 the breakpoint locates between exon 8 and exon 4. 9,37 In RS4;11 the KMT2A breakpoint has been shown to occur between exon 8 and 9 and between exon 5 and exon 4 in AFF1. 37,65 We confirmed the in-frame fusion of the two genes at these breakpoints generating the KMT2A-AFF1 transcript. Two distinct isoforms of the fusion are produced in patients with t(4;11), although their significance has not been elucidated. [66][67][68][69][70] The NAGNAG motif, identified on exon 7 of the AFF1 gene, is estimated to be present in 30% of human genes, and it was suggested that it may play a functional role in about 5% of the genes. 71 Analysis of an EST-derived alternative splicing database revealed that the NAGNAG motif is indeed subjected to alternative splicing in about 50% cases, 38 and moreover, the CAGCAG is a consensus sequence. 38,39 As the NAGNAG often undergoes alternative splicing, 38,39 the generation of noncanonical isoforms of the KMT2A-AFF1 transcripts is unsurprising.
The generation of the reciprocal fusion transcript AFF1-KMT2A from the der(4) was not investigated in our study, but its production has been reported in RS4;11. 72 The role of the AFF1-KMT2A protein in leukaemogenesis is still debated, as AFF1-KMT2A transcripts are only occasionally found in patients due to the fusion not consistently occurring in-frame. 73 While some authors have shown that AFF1-KMT2A is required to achieve full malignant transformation, 74,75 others did not find such association. 72,76 Although at least 16 cell lines with the t(4;11) have been described, only few have been appropriately authenticated (summarised inTable 2 ). 32,33 A similar cell line to RS4;11 is MV-4-11, harbouring a t(4;11)(q21; q23) together with an additional copy of chromosome 8 and chromosome 19. 84 MV-4-11 is morphologically macrophagocytic and was established from a 10-year-old male patient with biphenotypic myelomonocytic leukaemia. 46 While RS4;11 serves as a model for pre-B lymphoblastic leukaemia, MV-4-11 caters well for studies on myeloid and myelomonocytic cells. 33 The SEM cell line also carries a t(4;11), in conjunction with a del(7)(p14) and del(13)(q12), and was  Abbreviations: iAMP, intrachromosomal amplification; ITD, internal tandem duplication.