Challenging our understanding of B‐cell lymphomagenesis and risk: Paediatric high‐grade B‐cell lymphoma, not otherwise specified with a DDX3X::MLLT10 fusion and an IGH deletion

Abstract

We report a unique case of high‐grade B‐cell lymphoma, not otherwise specified in a 5‐year‐old child. Whole‐genome sequencing revealed a DDX3X::MLLT10 fusion, usually seen in T‐cell acute lymphoblastic leukaemia (ALL). This suggests the novel idea that MLLT10 fusions are capable of driving B‐cell malignancies. An IGH deletion usually only seen in adults was also found. These unique genetic findings provide novel insights into B‐cell lymphomagenesis. The child remains in remission 7 year post chemotherapy, which demonstrates that novel complex molecular findings do not always denote high‐risk disease.

Abbreviations

ALL

acute lymphoblastic leukaemia

AML

acute myeloid leukaemia

BL

Burkitt lymphoma

CLL

chronic lymphocytic leukaemia

DLBCL

diffuse large B‐cell lymphoma

HGBL

high‐grade B‐cell lymphoma

HGBL, NOS

high‐grade B‐cell lymphoma, not otherwise specified

LBCL

large B‐cell lymphoma

LBL

lymphoblastic lymphoma

NHL

non‐Hodgkin lymphoma

WGS

whole‐genome sequencing

1. INTRODUCTION

Non‐Hodgkin lymphoma (NHL) is a heterogenous group of malignancies that typically present as high‐grade extranodal disease, are managed with chemotherapy, and have excellent survival outcomes. ¹ The WHO classification currently distinguishes over 70 different types of B‐cell NHLs, but Burkitt lymphoma (BL) and diffuse large B‐cell lymphoma (DLBCL) account for the vast majority of paediatric cases. Rarer types of NHL that can occur in children include large B‐cell lymphoma (LBCL) with IRF4 rearrangement, high‐grade B‐cell lymphoma (HGBL) with 11q aberration, EBV‐positive DLBCL, ALK‐positive LBCL and plasmablastic lymphoma. ² In even rarer cases, NHLs are difficult to classify due to atypical presentation, histology, immunohistochemistry or genetics. This makes the optimal management and prognosis unclear. ³ We report the case of a 5‐year‐old with NHL that was difficult to classify. It possessed features of both mature and immature B‐cell NHL, and whole‐genome sequencing (WGS) revealed mutations not reported before in a childhood lymphoma. These provide novel insights into B‐cell lymphomagenesis.

2. CASE DESCRIPTION AND RESULTS

A 5‐year‐old male with no significant past medical history presented with rapidly enlarging masses over the right temple and parotid, plus lymphadenopathy. The rest of their systemic examination was normal, as were their peripheral blood counts and lactate dehydrogenase. Magnetic resonance imaging (MRI) and staging computed tomography (CT) revealed no additional lesions, including no lesions within the central nervous system, but bone marrow assessment revealed 45%–50% blasts. Histopathological and immunohistochemical assessment of the tumour revealed medium‐sized CD79a‐ and PAX5‐positive B cells with blastoid morphology, irregular rasinoid nuclei, small nucleoli or less frequently one large nucleolus, and an overall ‘starry sky’ appearance (Figure S1). Further immunohistochemical analyses identified other immature features consistent with the observed blastoid morphology, including CD10 positivity and only patchy CD20 positivity. However, mature B‐cell features were also present, including positive surface immunoglobulin, MUM1 and BCL6, strongly positive CD20 in focal areas of lymphoma, and negative terminal deoxynucleotidyl transferase (TdT) and CD34. In addition, Ki67 staining identified the proliferative fraction to be 60%, and limited genetic analyses, which were standard in the clinic at the time of diagnosis, demonstrated no rearrangements or copy number changes in MYC, BCL2 or BCL6. Overall, histopathological, immunohistochemical and genetic analysis of this child's tumour led to the rare diagnosis of stage IV HGBL, not otherwise specified (HGBL, NOS) with blastoid morphology. The child was managed with 7 months of chemotherapy for mature B‐cell NHL with bone marrow involvement following a FAB/LMB96 approach, and remains in remission 7 years post chemotherapy.

Tumour and matched germline WGS were performed for the 100,000 Genomes Project. This revealed no pertinent germline findings and a relatively low somatic mutation burden. ⁴ Mutational signature analysis identified the ‘clock‐like’ signatures 1 and 5 contributing to most of the mutation burden (Figure 1A and Figure S2). Given the low overall mutational frequency, these signatures likely represent random mutation acquisition during normal ageing/growing. ⁵ In total, four somatic events were deemed pertinent (Figure 1B,C). Of these, the DDX3X::MLLT10 translocation and IGH deletion have not been reported in a paediatric B‐cell lymphoma before.

FIGURE 1.

Whole‐genome sequencing results. (A) Somatic mutational signature analysis. Mathematical methods (decomposition by non‐negative least squares) were used to assess the top mutation signatures contributing to the overall mutation burden observed in the tumour. These mutation signatures have been derived using the analysis of large sequencing datasets (10,952 exomes and 1048 whole‐genomes from 40 distinct tumour types) in order to group patterns of relative contextual frequencies of different single nucleotide variants. (B and C) Somatic events. (B) Linear genome plot of global somatic structural variants (top track) and copy number aberrations (middle track). (C) Tabular summary of the four identified clinically pertinent somatic variants. VAF: variant allele frequency.

MLLT10 fusions, including with DDX3X, are found in T‐cell acute lymphoblastic leukaemia (ALL), ⁶ , ⁷ T‐cell lymphoblastic lymphoma (LBL) ⁸ and acute myeloid leukaemia (AML). ⁹ In this case, the translocation, t(10;X)(p12.31;p11.4), results in an in‐frame fusion of DDX3X exons 1–4 and MLLT10 exons 18–24 (Figure 2 and Figure S3). ⁴ Multiple breakpoints exist in T‐cell malignancies for DDX3X::MLLT10, including one similar to this case: exons 1−3 of DDX3X and exons 17−24 of MLLT10. ⁷ In all, the MLLT10 octapeptide motif‐leucine‐zipper domain is maintained, which is crucial for leukaemogenesis. ⁸ , ¹⁰ Other maintained features include the DDX3X nuclear export signal and MLLT10 nuclear localisation signal. ⁷ Moreover, we suggest the novel idea that MLLT10 rearrangements drive B‐cell malignancies too.

FIGURE 2.

DDX3X::MLLT10 gene fusion. Fusion‐specific PCR primers were designed to confirm the presence of the DDX3X::MLLT10 gene fusion detected in the patient's tumour sample by whole‐genome sequencing. An amplicon at the expected size of 780 bp was detected in the patient's tumour sample only. This amplicon was subsequently sequenced to confirm the genomic breakpoints of the fusion, which are demonstrated by the schematic. Sequencing demonstrated a somatically acquired translocation between the short arm of chromosome X (cytogenetic band p11.4) and the short arm of chromosome 10 (cytogenetic band p12.31). The genomic coordinates of the breakpoints (genome reference build 38) map to intron 4 of the DDX3X gene (transcript NM_001356) and intron 17 of the MLLT10 gene (transcript NM_004641). This translocation opposes exon 4 of DDX3X to exon 18 of MLLT10, leading to a DDX3X::MLLT10 gene fusion. Fluorescent sequencing analysis (sequence data show the negative strand in the opposite direction) confirms this fusion. Upon splicing and transcription, this fusion is predicted to maintain the amino acid read frame and result in a DDX3X::MLLT10 chimeric protein. Block arrows: exons; grey: untranslated region; blue: DDX3X coding exons; red: MLLT10 coding exons; grey connecting lines: introns; amino acid Ser: serine; arg: arginine; Gly: glycine; Pro: proline; Val: valine.

The second somatic change that has not been identified in paediatric lymphomas before is an IGH deletion. These occur in mature adult B‐cell malignancies, including chronic lymphocytic leukaemia (CLL) ¹¹ and DLBCL. ¹²

The other somatic events (Figure 1C) are not unexpected. KMT2D is a tumour suppressor that regulates homeobox genes and is important in B‐cell receptor signalling and apoptosis. ¹³ CDKN2A/CDKN2B are tumour suppressors with roles in the cell cycle, apoptosis and immune responses. ¹⁴ Inactivation of these genes is common in leukaemias and lymphomas, ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ including HGBL, NOS. ¹⁵ , ²⁰ Indeed, concurrent CDKN2A/CDKN2B deletions were seen in three of four ALL cases with DDX3X::MLLT10 fusions. ⁷

In summary, we report a HGBL, NOS with unique molecular findings. The tumour possessed a DDX3X::MLLT10 fusion usually only seen in T‐cell disease and an IGH deletion usually only seen in adults.

3. DISCUSSION

We present a case of paediatric HGBL, NOS with unusual histology and immunohistochemistry and previously undescribed genetic changes. It is highly unusual to see a paediatric B‐cell lymphoma with genetic events usually found in malignancies with other origins, such as T‐cell ALL and CLL. This tumour appears to have been driven by a DDX3X::MLLT10 fusion and an accompanying IGH deletion, both of which have not been seen before in paediatric B‐cell NHLs. When DDX3X::MLLT10 fusions are observed in ALL, they increase MLLT10 expression, leading to increased homeobox gene activity. This promotes cell cycle progression and drives leukaemogenesis. ⁸ , ⁹ Although DDX3X::MLLT10 translocations have not been reported in B‐cell malignancies before, DDX3X loss of function mutations have been in BL, ²¹ , ²² DLBCL, ²¹ HGBL, NOS ¹⁵ and CLL. ²³ In BL, these buffer MYC‐induced proteotoxic stress during early lymphomagenesis. ²¹ As DDX3X::MLLT10 translocations decrease DDX3X expression in T‐cell LBL, ⁸ a similar effect may have facilitated early lymphomagenesis in this case, although independently of MYC. However, the oncogenic and tumour suppressor functions of DDX3X require further characterisation. ⁷ IGH deletions are usually seen in CLL ¹¹ and DLBCL, ¹² and it is thought that they may activate an unknown oncogene at 14q. ¹¹ Together, these changes, seen individually in either immature or mature B‐cell malignancies, challenge our understanding of the genesis of the lymphoma seen in this case.

Difficult to classify B‐cell NHLs are currently allocated HGBL, NOS or double/triple hit lymphoma within the WHO classification. ² These are rare and biological courses are difficult to predict, but they are typically thought to have worse outcomes than other lymphomas. ³ , ¹⁵ Of note here, poor prognoses are also seen with: MLLT10 rearrangements in T‐cell ALL ⁶ , ⁷ and AML ⁹ ; IGH changes in CLL and prolymphocytic leukaemia ¹¹ ; KMT2D mutations in DLBCL ²⁴ and mantle cell lymphoma ¹⁶ ; and CDKN2A/CDKN2B deletions in DLBCL, ¹⁴ follicular lymphoma, ²⁵ cutaneous T‐cell lymphoma ¹⁸ and ALL. ¹⁹ , ²⁶ Moreover, there are few sufficiently comparable cases of paediatric HGBL, NOS in the literature, as well as limited available understanding of the genetic changes seen in the current case. In situations such as this, clinicians may consider whether intensified therapy is required and/or whether there is a place for the use of novel therapeutics, such as MAPK ²⁷ or KDM5 ²⁸ inhibitors for KMT2D mutations; CDK4/6 inhibitors for CDKN2A/CDKN2B deletions ¹⁹ ; and DOT1L inhibitors for MLLT10 changes. ⁷ However, the child remains disease‐free 7 years after chemotherapy alone. This cautions against overtreating patients with novel complex molecular findings due to assumptions that these confer high risk. Indeed, as WGS is not ubiquitous in clinical practice, we may have more sequencing data for unusual and/or aggressive cases. Therefore, we should presume neither that all findings are responsible for this phenotype, nor that these findings would not be found in typical disease.

In summary, paediatric HGBL, NOS is a rare entity that is difficult to diagnose, select treatment, and predict prognosis for. The child in this case had novel genetic changes, but remains disease‐free following chemotherapy alone. The case provides novel insights into paediatric B‐cell lymphomagenesis and highlights that NHLs are a more heterogenous group than the WHO classification currently accounts for. Indeed it is most likely that lymphoma is a continuous spectrum of disease, and the classification boundaries imposed are artificial. Therefore, a multi‐national effort to sequence these rare tumours is required to improve understanding of B‐cell lymphomagenesis, and to optimally categorise and manage these cases.

CONFLICT OF INTEREST STATEMENT

Lucy Hare, Jamie Trotman, Patrick Tarpey and Elizabeth Hook declare they have no conflicts of interest. G. A. Amos Burke has received institutional consultancy fees from Roche, Takeda, Novartis and Janssen.

FUNDING INFORMATION

Lucy Hare is supported with funding from a Cancer Research UK Cambridge Centre clinical research fellowship (grant number: C9685/A25117).

Supporting information

SUPPORTING INFORMATION FIGURE S1 Bone marrow smear histology. Histological appearances of the bone marrow smear at 100× (A) and 400× (B) magnification. Images were obtained using an Olympus BX51 microscope and Infinity Capture (Infinity 2‐1C) camera.

SUPPORTING INFORMATION FIGURE S2 Somatic mutational signature analysis. The counts of each mutation type (i.e., base substitution) at each mutation context (i.e., base situated immediately 3′ and 5′ to the mutated nucleotide) are normalised by the total number of tri‐nucleotide counts in the reference genome. All substitutions are referred to by the pyrimidine context of the mutated base pair. Mutation types are given on the horizontal, while the percentage of mutations attributed to a specific mutation type are on the vertical axis.

SUPPORTING INFORMATION FIGURE S3 PCR confirmation of the DDX3X::MLLT10 gene fusion detected in the patient's tumour sample. Fusion‐specific PCR primers were designed to confirm the presence of the DDX3X::MLLT10 gene fusion detected in the patient's tumour sample by whole‐genome sequencing. An amplicon at the expected size of 780 bp was detected in the patient's tumour sample only.

Hare L, Trotman J, Tarpey P, Hook E, Burke GAA. Challenging our understanding of B‐cell lymphomagenesis and risk: Paediatric high‐grade B‐cell lymphoma, not otherwise specified with a DDX3X::MLLT10 fusion and an IGH deletion. Pediatr Blood Cancer. 2024;71:e30810. 10.1002/pbc.30810