Rare MLL-ELL fusion transcripts in childhood acute myeloid leukemia—association with young age and myeloid sarcomas?

Background The chromosomal translocation t(11;19)(q23;p13) with a breakpoint within subband 19p13.1 is found mainly in acute myeloid leukemia (AML) and results in the MLL-ELL fusion gene. Variations in the structure of MLL-ELL seem to influence the leukemogenic potency of the fusion in vivo and may lie behind differences in clinical features. The number of cases reported so far is very limited and the addition of more information about MLL-ELL variants is essential if the possible clinical significance of rare fusions is to be determined. Case presentation Cytogenetic and molecular genetic analyses were done on the bone marrow cells of a 20-month-old boy with an unusual form of myelomonocytic AML with multiple myeloid sarcomas infiltrating bone and soft tissues. The G-banding analysis together with FISH yielded the karyotype 47,XY, +6,t(8;19;11)(q24;p13;q23). FISH analysis also demonstrated that MLL was split. RNA-sequencing showed that the translocation had generated an MLL-ELL chimera in which exon 9 of MLL (nt 4241 in sequence with accession number NM_005933.3) was fused to exon 6 of ELL (nt 817 in sequence with accession number NM_006532.3). RT-PCR together with Sanger sequencing verified the presence of the above-mentioned fusion transcript. Conclusions Based on our findings and information on a few previously reported patients, we speculate that young age, myelomonoblastic AML, and the presence of extramedullary disease may be typical of children with rare MLL-ELL fusion transcripts. Electronic supplementary material The online version of this article (doi:10.1186/s40164-016-0037-2) contains supplementary material, which is available to authorized users.

The leukemogenic potency of MLL-ELL fusion genes was demonstrated in murine model systems [17]. Moreover, variant forms of MLL-ELL were shown to impair transforming activities in vitro [12]. These observations suggest that variations in MLL-ELL structure may influence leukemogenic potency of the fusion also in vivo, and they hint that such variability may be behind variation in clinical features. Because so few such cases have been reported, the addition of more cases with MLL-ELL variants is essential if the possible clinical significance of rarer fusions is to be determined. In the present study, we report a childhood leukemia in which a three-way translocation caused the fusion of exon 9 of MLL with exon 6 of ELL. To the best of our knowledge, this is only the second case in which exon 6 of ELL was found to be fused to MLL [8,16,18].

Ethics statement
The study was approved by the regional ethics committee (Regional komité for medisinsk forskningsetikk Sør-Øst, Norge, http://helseforskning.etikkom.no), and written informed consent was obtained from the patient's parents to publication of the case details. The ethics committee's approval included a review of the consent procedure. All patient information has been anonymized.

Clinical presentation
A 20-month-old, previously healthy boy presented with intermittent pain in the hip. The boy's general condition was reduced, he was pale and weak, and passive movements in the right hip were painful. His spleen was slightly enlarged, whereas liver size was normal. Blood tests revealed anemia (hemoglobin 8.2 g/dL), thrombocytopenia (64 × 10 9 /L), normal white blood cell count (11.9 × 10 9 /L) with monocytosis (1.5 × 10 9 /L), and immature myeloid cells were seen in the blood smear ( Fig. 1).
A bone marrow investigation (which was repeated several times due to diagnostic difficulties) and a biopsy from one of the bony lesions revealed morphologically a myeloid hyperplasia comprised of predominantly monomorphonuclear cells. Erythropoiesis and megakaryopoiesis were sparse without dysplasia and decreased further with time until they were nearly absent in the later biopsies. The monomorphonuclear cells were positive for CD4 and CD56. The proportion of CD34+ and CD117+ cells was not increased. Further, flow cytometry analysis ( Fig. 2) confirmed the normal distribution of CD34 + or CD117 + precursor cells but revealed two distinct myeloid populations comprised of 68 % CD15 bright + /HLA-DR − granulocytic cells and 15 % CD64 bright + /HLA-DR − monocytic cells. Both lineages showed marker expression consistent with a shift to early stages as well as abnormal expression of CD56. In addition, CD13, a common myeloid marker, was not expressed in the early stages of these myeloid populations whereas abnormal expression of CD4 was demonstrated on early granulocytes. Only 50 % of the granulocytes had acquired expression of CD11b and to a lesser extent CD16 which is consistent with maturation beyond the promyelocyte stage. However, CD10 + /CD16 + end-stage neutrophils were not identified. Around 20 % of the monocytes were negative for CD35 and CD14 which are features of promonocytes. Based on the above immunophenotypic findings alone, a definitive diagnosis of acute myeloid leukemia could not be made.
Whole body magnetic resonance imaging (WBMRI) showed marked spiculated periosteal reaction and new bone formation in the mandible, iliac bones, and bilaterally in the distal femur ( Fig. 3a-d). In addition, there were widespread bone lesions with infarctions and necroses involving the skull base, mandible, scapulae, pelvic bones, femurs, and tibia. There was soft tissue involvement of the lower extremities with edema in the muscles and subcutaneous tissue and fluid along the facial planes ( Fig. 3a-d).
Mainly on the basis of the detected MLL-rearrangement, we interpreted the boy's disease as an unusual form of myelomonocytic AML with multiple myeloid sarcomas infiltrating bone and soft tissues.
During the investigation, the patient's extensive skeletal lesions increased causing therapy-resistant pain and his general condition deteriorated. After 3 weeks, AMLdirected therapy was begun according to the NOPHO-DBH AML 2012 protocol [19]. He received five courses (MEC, ADxE, HAM, HA3E, FLA) at 5-6 weeks intervals. For details, see Additional file 1: Figure: S1. Clinically, the boy recovered rather quickly. Evaluation before course two showed no remaining MLL-rearranged cells by FISH in the bone marrow. MRI controls of his bony lesions demonstrated continuous, but slow, regression. At 9 months after cessation of treatment, the boy is  (11), and der(19) of the t(8;19;11)(q24;p13;q23) together with the corresponding normal chromosome homologs; breakpoint positions are indicated by arrows. b FISH using an MLL breakapart probe showed rearrangement of MLL. The 3´-end part of the MLL gene (red probe) has moved to the q arm of the der(8), while the 5´-end part of the gene (green probe) remains on 11q23 of the der (11). c The initial RT-PCR amplifications for the detection of a possible MLL-ELL fusion transcript. Lane 1, nested PCR with the forward primers located in exon 7 of MLL and reverse primers located in exon 4 of ELL (MLL-3947F1/ELL-415R) failed to amplify any cDNA fragments. Lane 2, amplification of a cDNA fragment of the ABL1 gene using the primers ABL1-91F1 and AsBL1-404R1 suggested that the synthesized cDNA was of good quality. d RT-PCR using a new reverse primer located in exon 8 of ELL (primer ELL-1044R1) and a forward primer located in exon 7 of MLL (primer MLL-3878F) amplified a cDNA fragment. M, 1 Kb DNA ladder. e Partial sequence chromatogram of the amplified fragment using the primers MLL-3878F and ELL-1044R1 showing the junction of the MLL-ELL chimeric transcript clinically healthy. As expected, MRI still shows several small residual bone lesions undergoing regression. Unexpectedly, however, the patient has developed a new, small (about 5 %) clone in the bone marrow with a solitary 7q deletion. This may or may not represent an emerging secondary malignancy [20][21][22][23], and the situation is being monitored closely to see whether the clone expands and gives rise to hematologically recognizable disease.

Initial RT-PCR experiments
Total RNA, isolated from the patient's bone marrow at the time of diagnosis, was reverse-transcribed using iScript Advanced cDNA Synthesis Kit for RT-qPCR (Bio-Rad Laboratories, Oslo, Norway) and cDNA corresponding to 20 ng total RNA was used as template in PCR amplifications as previously described [24,25]. The initial RT-PCR with the primer set MLL-3878F/ELL-498R1 as well as nested PCR with the primers MLL-3947F1/ELL-415R (the sequences of the primers are listed in Table 1) failed to amplify any cDNA fragments (Fig. 4c). At the same time, use of the primer set ABL1-91F/ABL1-404R1 led to amplification of an ABL1 cDNA fragment suggesting that the synthesized cDNA was of good quality (Fig. 4c).

RNA-sequencing
Because of the negative RT-PCR results, the less than typical cytogenetic findings, and the clinical picture, three µg of the total RNA extracted from the patients' bone marrow at the time of diagnosis were subjected to high-throughput paired-end RNA-sequencing at the Norwegian Sequencing Centre, Oslo University Hospital (http://www.sequencing.uio.no/) as described elsewhere [24,26]. The raw sequencing data were subsequently analyzed using FusionCatcher which is a program designed to detect fusion genes from high throughput sequencing data [27]. More than 100 potential fusion transcripts were found (Additional file 2: Table: S1), among them an MLL-ELL in which exon 9 of MLL (nt 4241 in sequence with accession number NM_005933.3) was fused to exon 6 of ELL (nt 817 in sequence with accession number NM_006532. 3). No reciprocal ELL-MLL fusion transcript was found.
In order to verify the fusion obtained with Fusion-Catcher, we used the "grep" command utility [28] to search for expressions composed of 10 nt of MLL and 10 nt of ELL upstream and downstream of the fusion point, respectively (Table 2). Using the expression "GACTT TAAGGTGGCCAACAT" which is composed of 10 nt, "GACTTTAAGG", from MLL and 10 nt, "TGGC CAACAT", from ELL, 38 sequences were retrieved ( Table 2).

Molecular genetic confirmation of the fusion
PCR with the MLL-3878F and ELL-1044R1 primer combination (Table 1) amplified a fragment from the patient's bone marrow cDNA (Fig. 4d). Sanger sequencing of the amplified product showed that it was a chimeric MLL-ELL cDNA fragment in which exon 9 of MLL was fused to exon 6 of ELL, i.e., the same MLL-ELL fusion transcript found by RNA-sequencing ( Fig. 4e; Table 2).

Conclusions
We report a case of AML genetically characterized by a three-way translocation, t(8;19;11)(q24;p13;q23), leading to rearrangement of the MLL gene and the generation of a chimeric MLL-ELL transcript with fusion of MLL exon 9 to ELL exon 6. The initial RT-PCR amplifications relied on forward primers located in exon 7 of MLL and reverse primers located in exon 4 of ELL; this choice was based on findings in previous studies in which MLL was shown to fuse with exon 2 or 3 of ELL [3,10,11,15,29,30]. The PCRs with these primer sets (first PCR with MLL-3878F/ ELL-498R1, then nested PCR with the primers MLL-3947F1/ELL-415R) failed to amplify any cDNA fragments. It was a combination of three methods-banding  Table 2 Sequences retrieved with the «grep» command using the expression "GACTTTAAGGTGGCCAACAT" cytogenetics, FISH, and RNA-sequencing-that helped us identify the present MLL exon 9-ELL exon 6 fusion. G-banding analysis showed what appeared to be a regular chromosomal translocation t(11;19)(q23;p13) (Fig. 4a), a well-known change in acute leukemia. FISH showed that although MLL was split, the distal part of the gene was moved not to the derivative 19 but, surprisingly, to the long arm of chromosome 8 (Fig. 4b). Finally, RNAsequencing showed that exon 9 of MLL was fused to exon 6 of ELL (Table 2). RT-PCR using a new reverse primer located in exon 8 of ELL (primer ELL-1044R1, MLL-3878F and ELL-1044R1 primer combination) then confirmed the fusion transcript (Fig. 4d, e). In 2009, De Braekeleer and coworkers reported a case of congenital acute monoblastic leukemia with a threeway translocation, t(1;19;11)(p36;p13.11;q23), which involved the MLL gene and generated an MLL-ELL fusion identical to that of the present case [8,16,18]. To the best of our knowledge, these are the only two cases hitherto reported in which exon 6 of ELL is fused to MLL. It is certainly intriguing that three-way translocations, an unusual phenomenon behind MLL-rearrangements, had occurred in both cases; it may hint at some currently hidden mechanism behind the generation of the genomic change. The number of AML cases with MLL exon 9-ELL exon 6 fusions might actually be underestimated when assessed by means of RT-PCR amplifications using primer sets based on hitherto published studies [3,10,11,15,29,30]. This situation may be remedied by use of a new RT-PCR method that includes primers to detect fusion of MLL also with exon 6 of ELL [18].

CAGCACTCTCTCCAATGGCAATAGTTCTAAGCAAAAAATTCCAGCAGATGGAGTCCACAGGATCAGAGTGGACTTTAAGGTGGCCAACATGAGTGCTAAGGACG-
Current knowledge holds that there are three types of MLL-ELL fusion transcripts (Fig. 5). Type 1, found in the majority of cases, is characterized by ELL exon 2 being fused to 5´-MLL [3,[8][9][10][11][12][13]. Type 2 was so far found in only a single case of chronic myelomonocytic leukemia (CMML) that transformed to AML; it had exon 3 of ELL fused to 5´-MLL [15]. Type 3, of which the present case is one of two, is caused by a three-way translocation leading to the fusion of exon 6 of ELL with MLL (present case; [8,16,18]). All three MLL-ELL fusion proteins contain the aminoterminal region of MLL which includes the AT hooks, the methyltransferase domain, and the repression domain [31].
The case we describe presented unusual clinical features. Most conspicuous were the widespread, very painful bone lesions with soft tissue involvement which were interpreted as multiple myeloid sarcomas and extensive bone infarcts, not mere marrow infiltration. The validity of this diagnosis is supported by the gradual, protracted resolution of these lesions taking place during and after therapy. To the best of our knowledge, only four cases, including the present one, have been reported of very young AML patients displaying myelomonocytic features, myeloid sarcomas, and involvement of the MLL-ELL fusion gene [8,16,32,33]. The four patients have similar cytogenetic and genetic (MLL-ELL fusion) features. Three of them had a three-way translocation generating the MLL-ELL fusion: the present case with t(8;19;11)(q24;p13;q23), a female newborn with t(1;19;11) (p36;p13.11;q23) [16,32], and a three-month-old boy with t(6;19;11)(p22;p13;q23) [33]. None of the patients had the type 1 MLL-ELL fusion (see above). In the patient with t(1;19;11)(p36;p13.11;q23), the data suggest an MLL exon 9-ELL exon 6 fusion transcript, similar to our case [16,32]. In a two-month-old child reported by De Braekeler et al, the genomic breakpoints in MLL and The sequences of ELL are in bold Table 2 continued ELL indicated an MLL exon 9-ELL exon 3 fusion transcript [8,32]. In the three-month-old boy with t(6;19;11) (p22;p13;q23), the translocation resulted in an MLL exon 8-ELL exon 3 fusion transcript [33].

CTCAGCACTCTCTCCAATGGCAATAGTTCTAAGCAAAAAATTCCAGCAGATGGAGTCCACAGGATCAGAGTGGACTTTAAGGTGGCCAACATGAGTGCTAA-
Due to the small number of patients described, it is impossible to make definite statements about prognosis. Nevertheless, treatment results so far on patients carrying rare MLL-ELL fusion genes seem to have been encouraging. The newborn patient did not receive antileukemic therapy and died 24 h after birth [32], but the three treated patients, including the present case, went into remission [32,33]. Two of them seem to be long-term survivors [32,33], and our patient is in complete clinical remission 1 year after diagnosis. The clinical importance of the small clone with a 7q deletion that has emerged in remission is unclear. Recurrent cytogenetic abnormalities are sometimes seen in AML and ALL patients who are in complete clinical remission and may persist for years in the bone marrow even in the absence of progression to leukemia [20][21][22][23]. A wait-and-see approach is therefore prudent.