Introduction
Cancer is now accepted to be a genetic disease in the sense that it arises due to acquired genetic abnormalities in susceptible somatic cells (1). Microscopic studies of cancer cells have shown that these aberrations are often visible as balanced chromosomal changes, such as translocations and inversions, as well as unbalanced anomalies, such as deletions, monosomies, duplications, and trisomies (1). Many hematologic malignancies, including acute myeloid leukemia (AML) and acute lymphoblastic leukemia (ALL), are characterized by the presence of acquired chromosome translocations and inversions resulting in chimeric genes of pathogenetic, diagnostic, and prognostic importance (1). Whereas some genes, e.g., ABL, BCR, RUNX1T1, and PML, have only been reported involved in one or a few translocations, other genes are promiscuous, having numerous fusion partners in various translocations and even in different types of malignancy suggesting that the pathogenetic and phenotypic impact of the chimeras is dependent on both genes participating in the fusion (1).
One such gene is RUNX1 at 21q22 (2) which codes for the alpha subunit of the heterodimeric transcription factor named core binding factor (CBF) that binds to the core element of many enhancers and promoters. To date, RUNX1 (previously called AML1, CBFA2, PEBP2aB) has been shown in both myeloid and lymphoblastic acute leukemias to fuse with more than 30 different partner genes encoding a heterogeneous group of structurally diverse proteins (1). Recently, RUNX1 fusions were also found in adenocarcinoma of breast and lung as well as in squamous cell carcinoma of the oral cavity (3). Some of the fusions are common, such as ETV6-RUNX1 [t(12;21)(p13;q22)] in pre-B-ALL, RUNX1-RUNX1T1 [t(8;21) (q22;q22)] in AML, and RUNX1/MECOM [t(3;21)(q26;q22)] in myelodysplasia (MDS), AML, and chronic myeloid leukemia in blastic phase, whereas others have been reported in single cases, i.e., they have not yet been shown to be recurrent (2,4). The prognostic impact of the common RUNX1 fusions is well known (5–8). Corresponding knowledge for the infrequent RUNX1 chimeras is lacking (9).
Acquired point mutations distributed throughout RUNX1 are also frequently found in both de novo and secondary (therapy-related) MDS/AML (10,11). They are not found together with RUNX1 chromosomal translocations or complex abnormal karyotypes, and they are associated with poor outcome in MDS (12–16). The mutation spectrum includes missense, nonsense, frameshift, in-frame insertion/deletion mutations, as well as exon-skipping mutations (15). Nonsense mutations in RUNX1 account for 11% of the total and generate a repertoire of truncated RUNX1 proteins which to varying degree show lack of the C-terminal region. Most of them affect the transactivation domain (15).
Although less frequent, truncated RUNX1 proteins can also be the result of a chromosomal translocation which generates a premature stop codon in the RUNX1 open reading frame, leading to expression of C-terminal truncated forms. These chromosome translocations can be divided into two categories: in the first, the translocations produce only out-of-frame fusion transcripts (17–25) whereas, in the second category, they generate both in-frame and out-of-frame fusion transcripts (26–31).
The generation of C-terminally truncated RUNX1 proteins via different mechanisms suggests that their expression is important in leukemogenesis. Truncated RUNX1 protein was shown to reduce the transactivation capacity of CBF on specific myeloid promoters that function as inhibitors of normal RUNX1 (18–20). Recently, the truncated RUNX1 protein resulting from the t(1;21)(p32;q22) chromosomal translocation was shown to impair proliferation and differentiation of human hematopoietic progenitors (25).
Since acute leukemia treatment protocols are in part based on the presence of certain genetic changes, it is of clinical interest to obtain more information also about rare RUNX1 fusions, even in disease subgroups that so far cannot be treated with medications specifically directed against the leukemogenic defect. It is important to underscore that this may be the case also for infrequent pathogenetic mechanisms where information is gathered by the addition of single case reports, as recently exemplified by the story of the rare RUNX1-USP42 fusion and 5q deletion in AML (9,32–35).
For this reason, we here present the molecular genetic and clinical features of a case of AML with a cryptic t(6;21)(q25;q22) which resulted in the generation of a truncated RUNX1.
Patient and methods
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 de-identified.
Case report
A 7-year-old girl was admitted to the Children's Hospital because of petechiae. Prior to admission she had a one week history of fever, throat and abdominal pain and had been prescribed antibiotics on the suspicion of tonsillitis. On clinical examination, the girl was pale and had petechiae on the extremities and trunk, as well as a few hematomas on the legs. The peripheral blood values were hemoglobin 88 g/l, leukocytes 369.0×109/l, platelets 59×109/l, lactate dehydrogenase 1886 U/L, and C-reactive protein 71 mg/l. She had continuous epistaxis despite sustained platelet counts of 60×109 cells/l, normal international normalized ratio (INR), and activated partial thromboplastin time (APTT). There was no central nervous system involvement. Leucocytes gradually increased to 480.0×109 cells/l before start of the treatment.
Morphology and immunophenotypic findings were in keeping with the diagnosis acute myeloid leukemia with minimal differentiation (AML M0). Normal hematopoiesis was completely replaced by large blasts without conspicuous granulation or Auer rods and with lacy chromatin and prominent nucleoli. The blasts were positive for CD34, CD71, CD117, CD123, HLA-DR antigens, and the common myeloid markers CD13, CD33, and CD15. Less than 10% of the blasts were positive for cytoplasmic myeloperoxidase. Of interest, partial expression of Tdt and aberrant expression of CD7 and CD9 were demonstrated. The blasts were negative for B-cell, T/NK-cell as well as for monocytic, erythroid, and megakaryocytic lineage markers.
The bone marrow karyotype was 47,XX,+4[15] (see below). In addition, a FLT3 ITD mutation was detected, but no mutations in the nucleophosmin 1 gene. Upon induction treatment according to the NOPHO-AML 2004 protocol (NOPHO: Nordic Pediatric Hematology and Oncology) (36), morphologic remission (<5% blasts) was obtained. Due to the presence of a FLT3-ITD mutation, the patient became eligible for allogeneic stem cell transplantation (SCT). However, because a suitable donor was not found, consolidation therapy was completed with chemotherapy only. Four months after completed therapy, the patient had a bone marrow relapse. She went into a second remission on a clofarabin-based regimen and was transplanted with stem cells from her 7-month-old matching sibling. Unfortunately, she relapsed again 6 months after SCT and died one month later.
G-banding analysis
Bone marrow cells were cytogenetically investigated by standard methods. Chromosome preparations were made from metaphase cells of a 24-h culture, G-banded using Leishman stain, and karyotyped according to the ISCN 2009 guidelines (37).
Fluorescence in situ hybridization (FISH)
As part of our standard cytogenetic diagnosis, initial interphase FISH analyses of bone marrow cells were performed with the Cytocell multiprobe ALL panel (Cytocell, http://www.cytocell.co.uk/) looking for MYC rearrangements, CDKN2A (P16) deletion, TCF3 (E2A) rearrangements, ETV6-RUNX1 fusion, hyperdiploidy, MLL rearrangements, BCR-ABL1 fusion, and IGH rearrangements. On the basis of findings made using the above panel, further FISH was performed on metaphase spreads and interphase nuclei using the Vysis LSI TEL/AML1 ES Dual Color Translocation Probe (Abbott Molecular, http://www.abbottmolecular.com). This is a mixture of the LSI TEL probe labeled with SpectrumGreen and the LSI AML1 probe labeled with SpectrumOrange. Fluorescent signals were captured and analyzed using the CytoVision system (Leica Biosystems, Newcastle, UK).
RNA sequencing
Total RNA (3 μg) extracted from the patient's bone marrow at the time of diagnosis was sent to the Norwegian Sequencing Centre at Ullevål Hospital (http://www.sequencing.uio.no/) for high-throughput paired-end RNA-sequencing. The Illumina software pipeline was used to process image data into raw sequencing data. Only sequence reads marked as 'passed filtering' were used in the downstream data analysis. A total of 103 million reads were obtained. The FASTQC software was used for quality control of the raw sequence data (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). The software deFuse was used for the discovery of fusion transcripts (38) (http://compbio.bccrc.ca/software/defuse/).
In addition, the 'grep' command (http://en.wikipedia.org/wiki/Grep) was used to search the fastq files of the sequence data (http://en.wikipedia.org/wiki/FASTQ_format) for RUNX1 fusion sequences (NM_001754 version 4). To confirm the RUNX1 fusion identified by the deFuse program (see below), the 'expression' used was 'CAGATGCAGGAAGACTTTTG' which is a sequence of 20 nucleotides (nt) at the fusion point: 10 bases upstream (5′-end of RUNX1 gene, CAGATGCAGG), and 10 bases downstream from the junction (3′-end of the 6q25 intergenic sequence, AAGACTTTTG). The sequences obtained by 'grep' were blasted against the human genomic plus transcript database (http://blast.ncbi.nlm.nih.gov/Blast.cgi) as well as the reference sequences NM_001754 version 4 (RUNX1) and NC_000006.12 (chromosome 6).
PCR analysis
For reverse transcriptase-Polymerase Chain Reaction (RT-PCR), 1 μg of total RNA was reverse-transcribed in a 20 μl reaction volume using iScript Advanced cDNA Synthesis kit for RT-qPCR according to the manufacturer's instructions (Bio-Rad Laboratories, Oslo, Norway). The cDNA was diluted to 50 μl of which 1 μl was used as templates in subsequent PCR assays. The 25 μl PCR volume contained 12.5 μl Premix Ex Taq™ DNA Polymerase Hot Start Version (Takara Bio, AH diagnostics, Oslo, Norway), cDNA, and 0.4 μM of each of the forward and reverse primers. For detection of the RUNX1 fusion transcript, the forward RUNX1-809N-F1 (CGG CAG AAA CTA GAT GAT CAG ACC A) and reverse 6q25-R1 (TCC TTC AAG CAG CAA AAT CTG TGA G) primers were used. The PCR was run on a C-1000 Thermal cycler (Bio-Rad) with an initial denaturation at 94°C for 30 sec, followed by 35 cycles of 7 sec at 98°C, 30 sec at 60°C, 1 min at 72°C, and a final extension for 5 min at 72°C. PCR products (3 μl) were stained with GelRed (Biotium, Hayward, CA, USA), analyzed by electrophoresis through 1.0% agarose gel, and photographed. DNA gel electrophoresis was performed using lithium borate buffer (39). The remaining PCR products were purified using the GeneJET PCR Purification kit (Thermo Fisher Scientific, Oslo, Norway) and sequenced at GATC Biotech (Germany, http://www.gatc-biotech.com/en/home.html). The BLAST software (http://blast.ncbi.nlm.nih.gov/Blast.cgi) was used for computer analysis of sequence data.
Results
Cytogenetics
The G-banding analysis at diagnosis showed trisomy 4 in all 15 cells analyzed (Fig. 1A). The ETV6-RUNX1 probe showed abnormal signals with splitting of the RUNX1 probe in 203 out of 233 interphase nuclei examined in spite of no cytogenetically visible rearrangement of chromosome arm 21q (Fig. 1B). In the same experiment, 10 metaphase cells were examined in which part of the RUNX1 probe was unexpectedly seen to be located on the distal part of 6q (Fig. 1C). The data showed a novel cryptic t(6;21)(q25-27;q22) chromosome translocation (Fig. 1D). Other FISH analyses detected no rearrangements of MYC, TCF3, MLL, and IGH, no CDKN2A (P16) deletion, no hyperdiploidy, and none of the fusions ETV6-RUNX1 and BCR-ABL1. Therefore, the whole karyotype was: 47,XX,+4[15].nuc ish(ETV6x2,AML1x3) [209/233].ish t(6;21)(q25-27;q22)(AML1+;AML1+)[10] (Fig. 1A–D).
Analysis of RNA-sequencing with defuse
Using deFuse on the raw sequencing data, 39 potential fusion transcripts were found (data not shown), among them a fusion between RUNX1 and a sequence mapping close to the MIR1202 locus which corresponds well to the 6q breakpoint of the t(6;21) (q25-27;q22) suggested by combined G-banding and FISH. In order to verify the fusion obtained with the deFuse software, we used the 'grep' command utility to search for expressions composed of 10 nt of RUNX1 and 10 nt of 6q25 upstream and downstream of the fusion point (Table I). Using the expression 'CAGATGCAGGAAGACTTTTG', 9 sequences were retrieved which corresponded to the fusion RUNX1-transcript found by defuse (Table I).
Molecular confirmation of the RUNX1-fusions
PCR with the RUNX1-809N-F1/6q25-R1 primer combination amplified a 358 bp cDNA fragment (Fig. 1E). Direct sequencing of the amplified fragment verified the presence of the RUNX1-fusion transcript. The fusion point was identical to that found with deFuse (Fig. 1F). Therefore, the final karyotype after G-banding, FISH, and molecular examination could be written 47,XX,+4,t(6;21)(q25;q22)[10] (Fig. 1A and D).
Discussion
We present herein a case of childhood AML in which the leukemic cells had trisomy 4, a novel cryptic t(6;21)(q25;q22) chromosome translocation, and FLT3-ITD mutation. The molecular analysis of the translocation showed fusion of the RUNX1 gene with an intergenic sequence from 6q25 resulting in a putative RUNX1 truncated protein (Fig. 2A and B). The predicted truncated protein would contain the Runt homology domain (RHD) which is responsible for both heterodimerization with CBFB and DNA binding (40). Functionally, the truncated RUNX1 would be similar to the isoform AML1a of the RUNX1 protein (Fig. 2B, protein with accession number NP_001116079) (41–43). The isoform AML1a is a 250 amino acid RUNX1 protein which contains the RHD but lacks the proline-, serine-, and threonine-rich (PST) region which is the transcriptional activation domain at the C terminal end (41–43). AML1a does not itself have any transactivation function, but it inhibits the transcriptional activity of AML1b by competing for the DNA sequence of target genes with higher affinity (43). Overexpression of AML1a was shown to suppress granulocytic differentiation and to stimulate cell proliferation in 32Dcl3 murine myeloid cells treated with granulocyte colony-stimulating factor (43). AML1a was found to inhibit erythroid differentiation induced by sodium butyrate and enhance the megakaryocytic differentiation of K562 leukemia cells (44). AML1a also enhanced hematopoietic lineage commitment from human embryonic stem cells and inducible pluripotent stem cells (45). AML1a was reported to be highly abundant in the primitive stem/progenitor compartment of human cord blood, and forced expression of AML1a in these cells enhanced maintenance of primitive potential both in vitro and in vivo (46). Overexpression of AML1a was reported in patients with acute lymphoblastic leukemia and AML-M2 patients (47). In the same study, AML1a was found to repress transcription of promoter of macrophage colony-stimulating factor receptor mediated by AML1b (47). When murine bone marrow mononuclear cells were transduced with AML1a and then transplanted into lethally irradiated mice, the mice developed lymphoblastic leukemia after transplantation (47). Thus, AML1a seems to be an important contributing factor to leukemogenesis.
Truncated RUNX1 proteins generated by chromosomal translocations were shown to have functions similar to those of the AML1a isoform. In a patient with secondary AML carrying a t(19;21)(q13;q22), RUNX1 was fused out-of-frame to chromosome 19 sequences resulting in a truncated AML protein bearing the DNA binding domain but not the transcriptional activation domain. The fusion AML1 protein functioned as an inhibitor of the normal RUNX1 protein (19). The RUNX1-RPL22P1 (also known as AML1-EAP) fusion gene which is the result of the t(3;21)(q26;q22) chromosome translocation in AML, codes for a truncated RUNX1 protein which acts as an inhibitor of AML1b (17,18). The fusion of RUNX1 to CPNE8 in an AML with t(12;21)(q12;q22) also resulted in a truncated inhibitory RUNX1 protein (20). Recently, in vitro analysis of transduced human hematopoietic/progenitor stem cells showed that truncated RUNX1 proteins generated by a t(1;21)(p32;q22) chromosomal translocation increased proliferation and self-renewal and disrupted the differentiation program by interfering with AML1b (25). In a mouse model, truncated RUNX1 protein resulting from a point mutation induced pancytopenia with erythroid dysplasia, followed by progression to MDS-RAEB or MDS/AML (48). Dowdy et al studied the RUNX1 C-terminus in a mouse model by introducing a premature translational stop codon after amino acid 307 (Runx1Q307X) which mimicked RUNX1 mutations found in MDS/AML and CMML patients (49). They found that Runx1Q307X homozygous mice exhibited embryonic lethality at E12.5 due to central nervous system hemorrhage and a complete lack of hematopoietic stem cell function (49). They also showed that while the RUNX1 truncated protein was capable of binding to DNA, it was unable to associate with the nuclear matrix and failed to activate target gene promoters (49).
Taking all the above-mentioned data into consideration, it appears that the truncated RUNX1 protein (or absence from it of the C terminal part which contains subnuclear targeting and transactivation domains) is at least a contributing factor in leukemogenesis.
The patient described here also had, apart from the t(6;21)-RUNX1 rearrangement, trisomy 4 and FLT3-ITD mutation. The molecular genetic consequences of trisomy 4 are, as for numerical chromosome changes in general, unknown. Possible mechanisms could be global gene expression alterations because of gene dosage effect generated by the trisomy and duplication of any rearranged or mutated genes on chromosome 4. The prognosis for AML-patients with trisomy 4 is unclear, but based on a review of 30 such patients, Gupta et al (50) concluded that the outcome is poor compared to that of other cytogenetic subsets within the intermediate risk group. More importantly, a recent international collaborative study on pediatric t(8;21)-AML showed that gain of chromosome 4 in addition to t(8;21) represents a prognostically unfavorable feature (51).
FLT3-ITD mutation has been shown to be a prognostic factor although its impact has to be interpreted against the overall genetic background of the leukemic cells (52). In adult patients with a normal karyotype, FLT3-ITD is associated with poor prognosis (53,54). In core-binding factor (CBF) AML, higher mutant levels of FLT3-ITD were an adverse factor for overall survival (55). However, a recent report on adult patients with CBF AML stated that MRD levels, rather than the FLT3-ITD mutations, were significant prognostic markers for outcome (56). In pediatric patients, FLT3 mutations have been associated with poor prognosis (57,58). Reports on the significance of FLT3 mutations in pediatric CBF AML are lacking.
All in all, we cannot say which genetic event (+4, FLT3-ITD, or t(6;21)-RUNX1 truncation) was more pathogenetically or prognostically important. The case nevertheless illustrates that submicroscopic chromosomal rearrangements may accompany visible numerical changes and perhaps should be actively sought for whenever a single trisomy is found. To what extent and at which frequency such submicroscopic changes target the RUNX1 gene remains unknown. An active search for them may provide both pathogenetic and prognostic novel information in the future.