Apparent coexistence of ETV6::RUNX1 and KMT2A::MLLT3 fusions due to a nonproductive KMT2A rearrangement in B-ALL

Several recurrent disease-initiating genetic abnormalities define biologically, clinically, and prognostically distinct subsets of disease in B-lymphoblastic leukemia (B-ALL) [1]. Among these, B-ALL with t(12;21)(p13.2;q22.1) is characterized by the fusion of two transcription factors involved in the regulation of hematopoiesis, ETV6 and RUNX1, and is associated with a favorable outcome [2–4]. Conversely, B-ALL with t(v;11q23.3) harbors adverse prognostic implications and is defined by translocations involving KMT2A (formerly MLL) which encodes a histone methyltransferase involved in the regulation of gene transcription with one of several partner genes [5]. In line with the mutual exclusivity of primary genetic abnormalities in B-ALL [1,2,4], coexistence of t(12;21)(p13.2;q22.1) and t(v;11q23.3) is rare [6,7]. Herein, we describe the fourth case of B-ALL displaying apparent concurrent ETV6::RUNX1 and KMT2A fusions by fluorescence in situ hybridization (FISH). Considering the rarity of this constellation of genetic aberrations, we performed whole genome sequencing (WGS) to elucidate and characterize this case.

A three-year-old female presented to the emergency department with cough, rhinorrhea, and fever. Complete blood count revealed significant cytopenias with a hemoglobin level of 1.8 g/dL, platelet count of 4 × 10⁹/L and absolute neutrophil count of 0.220 × 10⁹/L. Total white blood cell count was 10.95 × 10⁹/L and the peripheral blood smear showed a blastic population. A bone marrow examination confirmed an infiltration by leukemic blasts accounting for 49% of cellular elements. By flow cytometry, these expressed CD45(dim), CD34(dim), HLA-DR, CD19, CD20, CD10, CD38, CD22, cCD22(dim), TdT, and cCD79a. Surface and cytoplasmic CD3, CD2, CD5, CD7, CD4, CD8, surface immunoglobulins, MPO, CD33, and CD13 were negative. These findings confirmed a diagnosis of B-ALL.

Conventional chromosome analysis was performed using standard cytogenetic techniques on white blood cells from the bone marrow aspirate. A FISH panel comprised of dual color, dual fusion probe sets (PBX1::TCF3, KMT2A::AFF1, KMT2A::MLLT4, KMT2A::MLLT3, KMT2A::MLLT10, KMT2A::ELL, KMT2A::MLLT1, ETV6::RUNX1 (laboratory-developed tests), ABL1::BCR (Abbott Molecular, Des Plaines, IL)), copy number probes (D4Z1, D10Z1, D17Z1, CDKN2A, D9Z1, TP53) and break-apart probes (MYC (Abbott Molecular), IGH (laboratory-developed test), CRLF2 and P2RY8 (Cytocell (OGT), Cambridge, UK)) was performed on cells from the bone marrow aspirate in accordance with specimen-specific laboratory protocols. Metaphase FISH was completed to confirm rearrangements of interest. Chromosomal microarray analysis (CMA) was performed on genomic DNA using the Global Diversity Array-8 (GDA) v1.0 Array BeadChip (Illumina, Inc., San Diego, CA) and analyzed using NxClinical (BioDiscovery, Inc., El Segundo, CA) software. WGS was performed on the Illumina NovaSeq 6000 sequencer using paired-end sequencing. Libraries were prepared using the modified NEB Ultra II (New England Biolabs, Ipswich, MA) and the Nextera Flex systems (Illumina, San Diego, CA). Reads from both libraries were combined and analyzed. FASTQ files were aligned to GRCh38 reference genome using BWA-MEM 0.7.17 [8]. Copy number alterations (CNAs) regions were calculated as any region that deviated from the expected diploid (2N) copy state by >10%. For structural variant (SV) detection, reads that mapped ≥5 Kb bp apart or to different chromosomes were considered discordant. Discordant fragments were clustered by fragment size and midpoint. Clusters from the normal peripheral blood WGS samples were used to create a 5 Kb mask to eliminate likely false positive results. Junction calls were also required to have ≥3 supporting fragments and a genomic footprint of ≥50 bp on both sides of the junction. SV and CNA calls were combined and visualized in a genome U-plot [9].

Institutional review board approval was obtained for this work and the patient gave written informed consent for research participation. Additional methods are found in supplementary materials. For original data, please contact [email protected].

Conventional chromosome analysis revealed a normal female karyotype (46,XX) in 15 metaphases and five additional metaphase cells were initially interpreted as having an abnormal complex karyotype with an apparent t(9;11)(p21;q23) and other chromosomal abnormalities (Figure 1(A)). Given the t(9;11) rearrangement, FISH studies using a dual color fusion probe for KMT2A::MLLT3 were performed and revealed KMT2A::MLLT3 fusion signals in ∼40% of interphase nuclei (Figures 1(B,C)) and in two metaphase cells (Figure 1(D)). Additional abnormalities were also observed including an ETV6::RUNX1 fusion in ∼42% of nuclei (Figure 1(E)), a deletion of CDKN2A (heterozygous in ∼15% and homozygous in ∼25% of nuclei) and trisomy 10 in ∼38% of nuclei. CMA revealed a 28.3 Mb interstitial deletion of 11q14.3q23.3 involving the 5′ end of KMT2A (exons 1 and 2) and a 12.6 Mb gain of 11q23.3q25 involving the 3′ region of KMT2A (from exon 3), suggesting that the breakpoint within KMT2A resided within an atypical location 5′ of the region between intron 9 and intron 11 where breakpoints involved in productive KMT2A fusions most commonly occur [5].

Figure 1. Conventional chromosome and FISH results. (A) Representative karyogram displaying a complex karyotype with a derivative chromosome 9 with inserted material from chromosomes 11 and X (arrow), monosomy 22, trisomy 10, and material of indeterminate origin at 11q13 and at 12p11.2. (B, C) Interphase nuclei demonstrating one fusion (B) and two fusion signals (C) involving KMT2A and MLLT3 per nucleus (arrow) with dual color, dual fusion FISH. (D) Metaphase nucleus illustrating derivative chromosomes 9 (black dashed arrow) and 11 (full black arrow) with a KMT2A::MLLT3 fusion and an uninvolved chromosome 11 (white arrow) with metaphase FISH. (E) Interphase nucleus demonstrating an ETV6::RUNX1 fusion (arrows) with dual color, dual fusion FISH.

Considering that dual-fusion FISH suggested an apparent KMT2A::MLLT3 fusion, yet findings with CMA appeared conflicting with those generally observed in KMT2A leukemogenic fusions, WGS was performed as part of a research study. This study confirmed the fusion involving ETV6 and RUNX1 (Figure 2(A)). While a rearrangement involving chromosomes 9 and 11 was also confirmed, WGS revealed that it juxtaposed KMT2A (from intron 1) with an intergenic region located 470 kb upstream of MLLT3. While this event did not result in an oncogenic fusion, it explained the dual fusion KMT2A::MLLT3 FISH results due to the close proximity of these genes (Figure 2(B)). WGS also confirmed homozygous and heterozygous loss of CDKN2A on chromosome 9p in different subclones within the population with an apparent KMT2A::MLLT3 fusion, suggesting that from an initial population harboring a heterozygous deletion involving CDKN2A on the derivative chromosome 9, two distinct subclones with homozygous loss of CDKN2A likely emerged. This homozygous loss was attributable to an additional deletion involving CDKN2A on the uninvolved chromosome 9 in the first subclone and to copy neutral loss of heterozygosity from endoreduplication of the derivative chromosome 9 in the second (Figure 2(C)). The re-analysis of the five metaphases incorporating the chromosomal microarray and WGS results revealed a 12.6 Mb insertion of 11q23.3q25 involving the 3′ region of KMT2A into chromosome 9p21.3 (46,XX,inv(5)(p13q31),der(9)ins(9;11)(p21;q23q25),der(9)ins(9;11)(p21;q23q25)t(X;9)(q21.31;q12), +10,del(11)(q14),add(12)(p11.2),-22 [5]/46,XX [15]). While the consequence of this insertion resulted in deletion of CDKN2A, it also positioned the KMT2A gene in close proximity to MLLT3 producing a misleading FISH result.

Figure 2. Schematic representation of genomic events involving KMT2A, MLLT3, CDKN2A, and ETV6::RUNX1 as identified by whole genome sequencing (WGS). (A) Representation of the ETV6::RUNX1 fusion [10]. (B1) Representation of the nonproductive KMT2A rearrangement resulting from an insertion of KMT2A (from intron 1) 470 kb upstream of MLLT3 and leading to the ‘false-positive’ fluorescence in situ hybridization (FISH) fusion signal. (B2) Genomic location of breakpoints on chromosomes 9 and 11. (C1) Illustration of normal chromosomes 9p and 11q with genomic regions labeled A through F, chromosomal microarray results and Normalized Read Depth (NRD) scores. (C2) Representation of the three likely subclonal malignant populations and the corresponding FISH signal patterns. From a clone alpha (α) with an insertion of a genomic sequence from chromosome 11 (including part of KMT2A) and a deletion involving CKDN2A on the derivative chromosome 9, two subclones (beta (β) and gamma (γ)) with homozygous loss of CDKN2A likely emerged (with copy neutral loss of heterozygosity (cnLOH) from endoreduplication of the derivative chromosome 9 in the β subclone (leading to the two apparent fusion signals by FISH) and from an additional deletion involving CDKN2A on the other chromosome 9 in the γ subclone). Schematic representations were created using ProteinPaint [10].

The patient was diagnosed with standard risk (COG) B-ALL without evidence of CNS involvement and received chemotherapy as per the AALL1731 COG protocol. Upon completion of induction chemotherapy, she achieved complete remission (CR) with minimal residual disease negativity by flow cytometry. At three months of follow-up, she currently remains in CR and is undergoing consolidation chemotherapy.

We describe the fourth case of B-ALL with apparent concurrent ETV6::RUNX1 and KMT2A fusions [6,7]. Our study reveals that the apparent fusion between KMT2A and MLLT3 by FISH was not predicted to result in a bona fide KMT2A::MLLT3 fusion and would not be expected to alter the intrinsic KMT2A regulatory mechanisms and confer leukemogenic potential [11]. Rather, it would be expected to result in a lack of protein production from this allele.

ETV6::RUNX1 and KMT2A fusions are considered disease-initiating primary genetic lesions in B-ALL with prevailing mutual exclusivity. Only three cases with apparent coexistent ETV6::RUNX1 and KMT2A fusions have been described in the literature. Interestingly, all cases had atypical findings by traditional cytogenetic investigation methods and shared demographic, pathological and prognostic features of ETV6::RUNX1-positive B-ALL [6,7]. Whether these cases also resulted from nonproductive KMT2A rearrangements remains conceivable yet uncertain as these were not investigated with high-resolution sequencing.

In summary, our report questions the possibility of co-occurrence of ETV6::RUNX1 and KMT2A fusions in the setting of B-ALL. Our case illustrates the importance of diligent consideration and interpretation of atypical results by current gold standard cytogenetic modalities, taking into account the expected structure of fusion products, as these modalities may sometimes yield clinically misleading results. Considering the diverging clinical implications associated with ETV6::RUNX1 and KMT2A fusions, WGS analysis was performed as part of a research study in the case we describe. However, many clinical laboratories are currently not performing WGS to resolve SVs and typically only rely on the gold standard cytogenetic modalities. In the setting of atypical findings with traditional cytogenetic methodologies, WGS may allow for characterization of genomic rearrangements and their putative oncogenic potential, ultimately heralding significant clinical implications.

Author contributions

L.B.B. designed the study. M.-F.G. and L.B.B. wrote the manuscript. J.B.S., N.S., and J.D.B. performed genomic analyses and analyzed the raw data. D.J.W. and J.J. provided patient care and clinical data. L.B.B, J.B.S., Y.M.N.A., J.D.B., R.M., M.-F.G., I.Z., D.J.W., C.A.S., P.T.G., X.X., N.L.H., R.P.K. and J.F.P interpreted the data. P.R.B., J.B.S., J.D.B., M.-F.G. and L.B.B. made the figures. All authors reviewed and approved the final manuscript.

Disclosure statement

The authors declare no competing financial interests.

Additional information

Funding

This work was supported by the Mayo Clinic Department of Laboratory Medicine and Pathology.