Molecular characterization and clinical outcome of B-cell precursor acute lymphoblastic leukemia with IG-MYC rearrangement

Abstract

Rarely, immunophenotypically immature B-cell precursor acute lymphoblastic leukemia (BCP-ALL) carries an immunoglobulin- MYC rearrangement (IG-MYC-r). This can result in diagnostic confusion with Burkitt lymphoma/leukemia and use of individualized treatment schedules of unproven efficacy. Here we compare the molecular characteristics of these conditions and investigate historic clinical outcome data. We identified 90 cases registered in a national BCP-ALL clinical trial/registry. When present, diagnostic material underwent cytogenetic, exome, methylome and transcriptome analyses. The outcomes analyzed were 3-year event-free survival and overall survival. IG-MYC-r was identified in diverse cytogenetic backgrounds, co-existing with either established BCP-ALL-specific abnormalities (high hyperdiploidy, n=3; KMT2A-rearrangement, n=6; iAMP21, n=1; BCR-ABL1, n=1); BCL2/BCL6-rearrangements (n=15); or, most commonly, as the only defining feature (n=64). Within this final group, precursor-like V(D)J breakpoints predominated (8/9) and KRAS mutations were common (5/11). DNA methylation identified a cluster of V(D)J-rearranged cases, clearly distinct from Burkitt leukemia/lymphoma. Children with IG-MYC-r within that subgroup had a 3-year event-free survival of 47% and overall survival of 60%, representing a high-risk BCP-ALL. To develop effective management strategies this group of patients must be allowed access to contemporary, minimal residual disease-adapted, prospective clinical trial protocols.

Introduction

Immunoglobulin (IG)-driven overexpression of the oncogene MYC is the genetic hallmark of mature, germinal center-derived Burkitt lymphoma (BL). However, IG-MYC translocations are observed in other mature B-cell malignancies including 5-14% of diffuse large B-cell lymphomas, high-grade B-cell lymphomas with BCL2-rearrangement (BCL2-r) and/or BCL6-rearrangement (BCL6-r), and multiple myeloma.^1-5 High MYC expression is driven by juxtaposition to powerful IG super-enhancers, most commonly the heavy chain locus resulting from the translocation, t(8;14)(q24;q32), but alternatively kappa or lambda light chain loci in the translocations t(2;8)(p11;q24) and t(8;22)(q24;q11), respectively.

Much less commonly, IG-MYC rearrangements (IG-MYC-r) have been identified in lymphoid malignancies expressing a surface immunoglobulin-negative, immature B-cell precursor (BCP) immunophenotype.^6-13 Several recent clinical trials of acute lymphoblastic leukemia (ALL) have excluded these patients to avoid the risk of mistreating mature B-cell non Hodgkin lymphomas (B-NHL). However, a large RNA sequencing study recently identified a distinct group of 18/1,988 (0.9%) BCP-ALL cases characterized by IG translocations with either BCL2, BCL6 and/or MYC,¹⁴ while a Nordic population-based study estimated the frequency of IG-MYC-r in childhood BCP-ALL at 0.6%.¹⁵ Despite the rarity of these conditions, knowing whether to diagnose and treat them as an ALL or Burkitt leukemia/lymphoma is important, with the relevant therapies being markedly different. A recent molecular study of 12 patients demonstrated genetic/epigenetic similarities to BCP-ALL.¹⁶ In contrast, a retrospective clinical study of 14 cases registered with the German BFM-NHL group encouraged treatment according to a mature B-NHL protocol.¹⁷ Clearly there is no current consensus, with many reported patients receiving hybrid protocols outside the clinical trial setting.^17,18

Here, we studied the molecular and clinical characteristics of a cohort of 90 BCP-ALL cases, enrolled in large national clinical trials/registries. Combining cytogenetic characteristics, somatic mutations, DNA methylation and gene expression analysis, we addressed the question of which disease precursor B-cell IG-MYC-r leukemias represent -BCP-ALL or BL. We further sought to determine the prognostic implications of the presence of an IG-MYC-r. We found that IG-MYC-r was present in diverse cytogenetic backgrounds and defined three subgroups within our cohort: (i) cases with IG-MYC-r and a cytogenetic abnormality recurrently seen in BCP-ALL; (ii) cases with IG-MYC-r and BCL2-r/BCL6-r; and (iii) the majority of cases, with IG-MYC-r as the defining cytogenetic abnormality. We demonstrate that IG-MYC-r can be either the candidate driver event or a secondary cytogenetic feature. We provide further evidence that the majority are distinct from BL and instead represent a subtype of BCP-ALL with a high risk of early relapse. To improve the care of these patients, clinical outcomes should now be assessed in the setting of prospective minimal residual disease-adapted clinical trials.

Methods

Sample cohort

Patients were identified by the UK Leukaemia Research Cytogenetics Group (LRCG), US Children’s Oncology Group (COG), Dutch Children’s Oncology Group (DCOG) or International Berlin-Frankfurt-Münster (BFM) study group members. The patients’ full characteristics, immunophenotype and additional methodological details are included in the Online Supplementary Data, Online Supplementary Tables S1 and S2 and Online Supplementary Figure S1. The primary inclusion criteria were enrollment/registration in a BCP-ALL clinical trial/registry and IG-MYC-r detected by karyotyping or fluorescence in situ hybridization (FISH) as a translocation involving an immunoglobulin locus (IG) and MYC: IGH-MYC-r, t(8;14)(q24;q32); IGK-MYC-r, t(2;8)(p11;q24); IGL-MYC-r, t(8;22)(q24;q11). We also identified two patients with T-cell receptor (TCR) translocations to the MYC locus. Non-trial cases were identified from local/national registries based on a local diagnostic immunophenotype of B-cell precursor malignancy. Children were defined as those patients under the age of 18 years at diagnosis. Informed consent and institutional review board approval were obtained at each collaborating center. This study was performed in accordance with the Declaration of Helsinki.

Conventional cytogenetics and fluorescence in situ hybridization

Fixed cells were available for 65 MYC-rearranged patients (Online Supplementary Table S1). IGH, IGK, IGL, MYC, BCL2 and BCL6 FISH studies were performed in all cases with sufficient available cells, in that order of priority. The labeling, capture and scoring methods are described in the Online Supplementary Methods.

Multiplex ligation-dependent probe amplification

Copy number alterations were determined by multiplex ligation-dependent probe amplification (MLPA) using the SALSA MLPA kit P335 (MRC Holland, Amsterdam, the Netherlands) in 27 patients.¹⁹ The kit includes probes to IKZF1, CDKN2A/B, PAX5, EBF1, ETV6, BTG1, RB1 and CSF2RA/IL3RA/CRLF2.

Targeted sequencing

DNA from 17 patients was prepared and sequenced using a targeted next-generation approach to identify translocation breakpoints at IG and MYC loci, as described previously.⁴ BWA-MEM (v0.7.12) and genome assembly GRCh37 was used for sequence mapping. Chromosomal rearrangements were called using Manta (v.0.29.6)^4,20 and manually inspected in Integrative Genomics Viewer (Broad Institute, Cambridge, MA, USA).

Whole exome sequencing

DNA libraries of 15 diagnostic and two relapse samples were generated using the TWIST Human Core Exome kit (Twist Biosciences, San Francisco, CA, USA) and paired end sequencing performed on a NovaSeq (Illumina, San Diego, CA, USA) at the Newcastle University Genomics Core Facility. Data were analyzed using the Genome Analysis Toolkit (GATK) including the variant caller MuTect2^21-23 (Broad Institute).

Illumina Infinium MethylationEPIC BeadChip array

Methylation data were generated for 18 diagnostic samples at Eurofins Genomics (Ebersberg, Germany) using Human-Methylation Epic arrays (Illumina) and the manufacturer’s standard protocols. To assess relationships between IG-MYC-r leukemia and other disease entities, we selected the 1,404 CpG loci previously reported to be related to B-cell maturation¹⁶ and performed t-distributed stochastic neighbor embedding (t-SNE) visualization of the combined sample set (Online Supplementary Table S3), as previously reported.²⁴

RNA-sequencing

RNA-sequencing was performed for six patients’ samples at Eurofins Genomics (Ebersburg, Germany) using a TruSeq Stranded mRNA Library Prep Kit (Illumina). Data were quantified using Kallisto²⁵ and analyzed using DEseq2.²⁶ These data were compared with publicly available data (EGAS00001001795).²⁷ The top 5% of the genes with the highest median absolute deviation were selected. The resulting count matrix was used for t-SNE analysis.²⁸

Outcome analysis

Event-free survival (EFS) was defined as time from diagnosis to relapse, second tumor, or death, censoring at last contact. Overall survival (OS) was defined as time from diagnosis to death, censoring at last contact. For patients without data on relapse or second tumor, death was assumed to be the first event. Survival rates were calculated and compared using Kaplan-Meier methods, log-rank tests, and Cox regression models (univariate and multivariate analyses). Other comparisons were performed using the χ² or Fisher exact test. All tests were conducted at the 5% significance level. All outcome analyses were performed using Intercooled Stata 15.0.

Results

Demographic and clinical characterization

We collected 90 cases of BCP-ALL diagnosed by local hematologists between August 1989 and July 2019 and subsequently identified as having an IG-MYC-r by karyotype or FISH (Online Supplementary Figure S1, Online Supplementary Table S1). The majority were registered at diagnosis in a BCP-ALL clinical trial/registry via the UKALL group (n=32), COG (n=32) or DCOG (n=5). Eighteen additional cases were identified by international BFM study group centers, and eight of these were enrolled on ALL trials. Three further patients in whom IG-MYC-r was only demonstrated at relapse were not included in the survival analysis. Patients received a wide variety of trial- and clinician-determined treatments.

Among 87 cases identified at presentation, the median age at diagnosis was 10 years (range, 0-81 years) (Table 1). Fifty-two (60%) patients were male, 35 (40%) female. The median presenting white blood cell count was 14x10⁹/L (range, 1.6-1200x10⁹/L, n=77 patients). Diagnostic bone marrow blast percentage was available for 64 patients and ranged from 30%-100% with a median of 86.5%. Central nervous system disease was present in 15/64 (23%) cases. Immunophenotype data were available for 33 diagnostic cases and one relapse case (Online Supplementary Table S2). TdT was positive in 22/29 (76%) cases. Of the TdT-negative cases, all seven (100%) were negative for surface immunoglobulin, confirming a BCP immunophenotype.

Table 1.Clinical characteristics of patients presenting at diagnosis with B-cell precursor acute lymphoblastic leukemia carrying an IG-MYC-rearrangement.

Cytogenetic characterization identified both heterogeneity and distinct genomic features

Karyotype was available for 89/90 patients (86/87 at diagnosis) and was abnormal in all cases. Three patients had high hyperdiploidy (51-65 chromosomes), a favorable risk feature and eight had high-risk genetics (KMT2A-r, n=6; iAMP21, n=1; BCR-ABL1, n=1) (Table 1, Online Supplementary Table S1). The remaining 75/86 (87%) cases were classified as intermediate cytogenetic risk at diagnosis. Fourteen patients (5 children, 9 adults) had a co-existing BCL2-r, one of whom also had a co-existing BCL6-r (Online Supplementary Table S1). In addition, one adult patient had only a BCL6-r co-existent with their IGH-MYC-r, identified by FISH. In the remaining 64 patients the IG-MYC-r was the defining cytogenetic feature. Karyotyping and/or FISH showed that 47/90 patients (52%) had rearrangements of the IGH locus, 39 (43%) the IGL locus and three (3%) the IGK locus (Figure 1A). Two (2%) patients had TCR-MYC-r. One patient had both IGH-MYC and IGL-MYC rearrangements.

We estimated the percentage of cells carrying IG-MYC-r in 53 patients with quantitative FISH data and found that, irrespective of the IG locus involved, IG-MYC-r can be either clonal or subclonal (Figure 1B, Online Supplementary Figure S2). Additional evidence for the subclonal nature of some IG-MYC-r was derived from 3/5 KMT2A-rearranged cases (22901, 30611, 17659) in which karyotype and FISH data demonstrated the presence of KMT2A-rearranged but MYC-germline cells, implying that the KMT2A-rearranged transformation preceded the MYC-r (Figure 1C, Online Supplementary Table S1). Furthermore, it was possible to reconstruct the clonal evolution of KMT2A-rearranged case 17659 from the presentation and relapse metaphase data (Figure 1D, Online Supplementary Table S1), showing that the relapse was derived from a clone carrying the cooperating IGL-MYC-r. Patients with other established ALL-specific cytogenetic abnormalities (high hyperdiploidy, iAMP21) harbored MYC-r within the same clone (Online Supplementary Table S1). Together these data show that, for at least a proportion of cases, IG-MYC-r are secondary chromosomal events. Nevertheless, when compared with MYC-germline cases of BCP-ALL, a high level of MYC expression was identified, irrespective of the IG partner or degree of clonality (Figure 1E). In nineteen patients, IG-MYC-r was the sole abnormality (Online Supplementary Table S1). Among the remaining cases, abnormalities of chromosome 1 were the most frequently observed aneuploidy/structural gain (28/89, 31%) (Figures 1F and 2, Online Supplementary Table S1). In the majority (25/28, 89%), 1q abnormalities were shown to be present within the same clone as the IG-MYC-r. One specific feature we identified, which has not been recurrently described in either BL^29-32 or BCP-ALL,³³ was isochromo-some of the long arm of chromosome 1, i(1)(q10). This was observed in nine patients and was associated with IGHMYC in eight.

Figure 1.Cytogenetic characterization of IG-MYC-rearranged patients. (A) Karyotype reveals the distribution of immunoglobulin chain involvement in MYC translocations (fluorescence in situ hybridization [FISH] data were used for one patient who presented with a normal karyotype). +BCL2/6-r: concomitant BCL2/BCL6-rearrangement; +ALL-r, concomitant established acute lymphoblastic leukemia rearrangement; IG-MYC: IG-MYC rearrangement as the defining cytogenetic abnormality. (B) Percentage of nuclei carrying IG-MYC-r grouped by immunoglobulin chain involvement. Blue dots, IG-MYC-r alone. Green dots, IG-MYC-r and established ALL rearrangement. Red dots, IG-MYC-r and BCL2/BCL6-r. Dotted and dashed lines, interquartile ranges. (C) Percentage rearrangement of MYC (blue dots) and KMT2A (green dots) using FISH. (D) Evolution of t(4;11) and subsequent gain of t(8;22) in case 17659. Percentages at diagnosis represent the proportion of metaphases seen with each abnormality (left panel). Representative chromosomes taken from diagnostic and relapse metaphases (right panel). Gray arrows mark the portion of chromosome 22 translocated to chromosome 8. (E) RNA sequencing comparing expression of MYC among (labeled) cases from the IG-MYC-r cohort with other in-house B-cell precursor ALL cases. TPM: transcripts per million reads. (F) Chromosomal abnormalities observed for chromosome 1 in the karyotypes of patients with IG-MYC-r. Red line: patients with concomitant BCL2-r.

We further analyzed copy number alterations in nine of the most commonly deleted genes/regions in ALL (Figure 3) for 27 patients with sufficient DNA. Among the 15 cases with IG-MYC-r as the defining cytogenetic abnormality, 11 (73%) showed no deletion within the genes/regions analyzed. This is a remarkably high percentage and even more so when specifically considering pediatric patients among whom 9/10 (90%) had no deletions at diagnosis. This contrasts with just 26% of childhood ALL classified as Bother, the subgroup to which all ten patients would be assigned,¹⁹ implying an important driver role for IG-MYC-r in this subgroup of patients.

Finally, in order to investigate the IG and MYC breakpoints at high resolution, we used a targeted sequencing approach⁴ and were able to resolve exact breakpoints for 15/17 patients analyzed (Figures 3 and 4, Online Supplementary Table S4). MYC breakpoints differed according to IG partner locus. IGH-MYC breakpoints (n=9) were detected 5’ of MYC, within the 5’ untranslated region or within intron 1. One case had an additional breakpoint within intron 2 (Figure 4A). Each resulted in MYC being juxtaposed to the IGH super-enhancers on chromosome 14. In contrast, all IGL-MYC breakpoints (n=5) resulted in the translocation of the IGL super-enhancer 3’ (telomeric) of MYC (Figure 4A). In patients with IGH-MYC-r, the majority (7/9, 78%) had breakpoints located with the V(D)J gene segments. Notably, just two patients had constant region breakpoints, more commonly associated with germinal center class switch recombination activity (Figure 4B). One of these cases also carried an IGH-BCL2-r (30279). In addition, this analysis identified three patients (26683, 27424 and 29785) who had cryptic IGH-DUX4-r,^27,34-36 following which they were considered within the group of cases with an established ALL-specific cytogenetic abnormality (Online Supplementary Table S1).

Together these analyses show significant variability in the cytogenetic characteristics of BCP-ALL with IG-MYC-r. In some cases the IG-MYC recombination is clearly a secondary chromosomal event but is still capable of imparting a potentially important biological effect. In other cases it is accompanied by additional IG translocations, notably to BCL2 and/or BCL6. However, in the majority of this cohort, IG-MYC-r occurred either in isolation or in association with additional cytogenetic features not typical of either BCP-ALL or BL. We therefore sought to investigate the relationship between these disease entities further.

Figure 2.Ideogram of cytogenetic rearrangements reported in the karyotypes of 70 patients with IG-MYC-rearrangements.

Figure 3.Oncoplot depicting the incidence of selected genomic features of IG-MYC-rearranged patients with material available for genomic studies. D: diagnosis; R: relapse; dark-shaded square: positive result; light-shaded square: negative result; white square: not tested; T: 1q translocation.

IG-MYC-rearranged B-cell precursor acute lymphoblastic leukemias display diverse genetic, epigenetic and transcriptomic features

In order to better characterize the molecular landscape of IG-MYC-r BCP-ALL we analyzed exome sequencing data from 15 diagnostic and two relapse samples, interrogating a panel of genes recurrently mutated in BCP-ALL or BL (Figure 3).^14,37-39 In keeping with a recent analysis of exome data from five similar cases,¹⁶ we found non-synonymous variants in KRAS in 5/15 (33%) patients at diagnosis. Mutations were also recurrently identified in genes mutated in BCP-ALL including, FAT4, ASMTL and IKZF1, observed in 6/15 (40%), 3/15 (20%) and 3/15 (20%) patients, respectively. Additional mutations were observed in ANK3, FAT1, FAT2 (2/15, 13%) and PDGFRA (1/15, 7%). However, the cohort also included two patients (4352 and 30279) who harbored BL hotspot mutations in ID3 (L64F) and TCF3 (N554K) (Online Supplementary Figure S3), respectively, and additional characteristic BL mutations in DDX3X, SMARCA4, and CCND3. Interestingly, case 30279 also had IGH-BCL2 and BCL6 rearrangements as have recently been described in a cohort of high-grade, molecular BL.⁴⁰ Wagener et al. were also able to analyze the methylome of two cases, neither of which carried an IG-BCL2/BCL6 rearrangement, and found these to be distinct from that of BL, instead clustering with BCP-ALL cases.¹⁶ Here, we integrated 18 of our cases with published DNA methyl-ation data of BCP-ALL^41-43 and BL^16,44 cases and identified four clusters (Figure 5A). The biggest, in which no IG-BCL2/BCL6 rearrangement or ALL-specific cytogenetic abnormality was identified (cluster A), contained seven cases (4 with IGH-MYC-r, 2 with IGL-MYC-r and 1 with TRA-MYC-r) and appeared most closely associated to BCP-ALL with the TCF3-PBX1 fusion gene. Distinct from cluster A, three patients clustered with cell lines and primary samples from BL patients (cluster B). These three patients carried either BL-like mutations in ID3/TCF and IGH constant region breakpoints (30279 and 4352) and/or IG-BCL2-r (30279 and 25729). Cluster C contained three cases identified by targeted IG-sequencing to have IGH-DUX4. Patient 30611 had a KMT2A-r in 40% of nuclei and clustered with other KMT2A-rearranged samples.

Furthermore, transcriptome analysis was conducted for six patients (5 with methylation data) for whom RNA was available (Figure 5B, C). Patients with IG-MYC-r and no IG-BCL2/BCL6-r or ALL-specific cytogenetic abnormality again associated, although here they were less close to BCP-ALL samples with TCF3-PBX1. Two IGH-DUX4 patients (27424 and 29785) were confirmed to cluster with published IGH-DUX4 samples, despite substantially different MYC-r frequencies of 16% and 90%, respectively. KMT2A-r patient 30611 associated with published KMT2A-r patients.

Figure 4.Targeted IG and MYC sequencing identified heterogeneous breakpoints. (A) Distribution of breakpoints within the MYC locus. (B) Distribution of breakpoints within the IGH locus. Each line shows the breakpoints for patients with translocations involving individual genomic loci. Frequency distribution defines a region of increased frequency of breaks (peach shaded area). Upper panels provide an expanded view of the breakpoint hotspots. Each dot represents an individual breakpoint. Blue dots: IGMYC-rearrangement alone. Green dots, IG-MYC-rearrangement and established acute lymphoblastic leukemia rearrangement. Red dots, IG-MYC-rearrangement and BCL2/BCL6-rearrangement.

Overall these analyses show that the four cases with an ALL-specific cytogenetic abnormality clustered with the relevant leukemic subgroup. In our methylome analysis, cases with either an additional BCL2/BCL6-r, IGH constant region break or a mutational profile in keeping with BL, clustered with BL. However, cases without BCL2/BCL6-r or a BL-like profile formed a distinct cluster, strongly indicating they are a subtype of BCP-ALL.

Outcome of patients with IG-MYC-rearranged B-cell precursor acute lymphoblastic leukemia

Our genomic analysis identified three groups of IG-MYC-r cases, defined by: (i) concomitant BCL2/BCL6-r; (ii) co-occurrence of an established ALL-specific cytogenetic abnormality; and (iii) IG-MYC-r in the absence of another defining cytogenetic abnormality (Figure 6). That analysis demonstrated at least a proportion of IG-MYC/IG- BCL2/BCL6-r “double/triple hit” cases to be biologically distinct from BCP-ALL, instead showing overlap with molecular BL as has recently been described.⁴⁰ In keeping with published case reports,^45,46 we identified a markedly inferior EFS (hazard ratio [HR]=3.10, 95% confidence interval [95% CI]: 1.42-6.76, P=0.004) and OS (HR=3.78, 95% CI: 1.63-8.79, P=0.002) in comparison to BCL2 wildtype cases. We therefore excluded these cases from subsequent analyses.

Figure 5.Methylation and transcriptome data identified five clusters within the cohort of patients with IG-MYC-rearrangements. (A) Consensus clustering of methylation data. Common t-distributed stochastic neighbor embedding (t-SNE) visualization of IGMYC-rearranged patients combined with publicly available data for subtypes of B-acute lymphoblastic leukemia (GSE49031, GSE69229, GSE76585), a patient with Burkitt lymphoma (GSE114210) and cell line samples (GSE92378). Each dot represents an individual patient. (B) Consensus clustering of RNA-sequencing data. Common t-SNE visualization of IG-MYC-rearranged patients combined with publicly available RNA-sequencing data for subtypes of B-acute lymphoblastic leukemia (EGAS00001001795). Each dot represents an individual patient. (C) Patients’ identities assigned to clusters. *Methylation and RNA-sequencing data, **RNA-sequencing only.

For those cases with an ALL-specific cytogenetic abnormality, the IG-MYC-r was often subclonal and DNA methylation/transcriptomic analysis confirmed that they continued to cluster according to the underlying cytogenetic subgrouping. We believe that these cases present less of a diagnostic challenge and should continue to be eligible for inclusion in ALL clinical trials. We therefore further investigated the clinical course of the diagnostically more challenging patients in whom IG-MYC-r represented the only recurrent cytogenetic feature. Among 59 cases enrolled in an ALL clinical trial/registry, follow-up data were available from 55 patients. However, only 20 patients (15 children and 5 adults) received the full protocol on trial (Online Supplementary Figure S1). The remaining patients received unknown off-trial treatment, which was at the clinical team’s discretion. As a result, these off-trial patients had markedly shorter follow-up available, particularly with regard to relapse events (Figure 7). Within the small subcohort suitable for analysis, we did not identify a survival difference between children treated on an ALL trial versus those taken off an ALL trial with 3-year EFS being 47% (21-69) versus 67% (37-85) (P= 0.13) and OS being 60% (32-80) versus 80% (50-93) (P=0.15), respectively. Although there were long-term survivors among trial and non-trial patients, the predominant feature among both children and adults was a high frequency of early relapse, consistent with high-risk disease (Figure 7A, B). Insufficient minimal residual disease data were available to allow analysis.

Twenty-six children demonstrated a molecular feature identified within our study (IG-MYC-r clonality >50%, KRAS mutation, abnormal chromosome 1 or i(1)(q10) (Online Supplementary Figure S4). Univariate analysis did not identify any of these to be associated with EFS or OS (Online Supplementary Table S5).

Figure 6.Molecular characteristics of cases of IG-MYC-rearranged acute lymphoblastic leukemia. Summary of three groups defined within the molecular analysis. Cases with a co-existing acute lymphoblastic leukemia (ALL)-specific cytogenetic abnormality are shown in green in the left panel. Cases with an IG-MYC-rearrangement (IG-MYC-r) as the only recurrent cytogenetic feature are shown in blue in the center panel and cases with a co-existing rearrangement of BCL2 and/or BCL6 are shown in red in the right panel. Each donut plot shows the proportion of cases studied which have a specific molecular feature. The total number of cases in each study is shown within each donut. The upper donut plots show the proportion with variable V(D)J gene segment breakpoints (dark segments) and constant segment breakpoints (light segments). The middle donut plots show the proportion of cases with either Burkitt-like TCF3/ID3 mutations (light segments), KRAS mutations (medium segments) or neither such mutation (dark segments). The lower donut plots show the proportion of cases within each of three DNA methylation clusters, either Burkitt-like cluster B (light segments), the IG-MYC-r cluster A (medium segments) or one of the other ALL specific clusters (dark segments). *The single case in the center panel harboring both a constant region breakpoint and an ID3 mutation (case 4352) clustered with Burkitt lymphoma/leukemia cases in our DNA methylation analysis.

Discussion

Although rare, the identification of an IG-MYC-r in the karyotype of patients diagnosed with BCP-ALL challenges diagnostic and clinical decision-making, with many case reports/series demonstrating the lack of uniform management, often involving hybrid ALL/B-NHL treatment.^17,18,46 Furthermore, given the poor outcome of non-lymphoblastic/Burkitt leukemia treated with ALL therapy,⁴⁷, ^{4 8} several ALL clinical trials have excluded patients with IG-MYC-r, restricting the development of effective protocols. To address this clinical need we have collected the largest published cohort of IG-MYC-r BCP-ALL cases, from the UK, Europe and the USA, studying their molecular biology and clinical outcome.

The principal molecular distinction required is from Burkitt leukemia/lymphoma, which shares the hallmark cyto-genetic feature of an IG-MYC-r. However, the majority (approximately 80%) of BL cases have an IGH-MYC-r, while the remainder have IGL or IGK rearrangements in similar proportions.⁴⁹ In contrast, we identified a much higher frequency of light chain rearrangements and particularly IGL-MYC-r. A substantial proportion of BCP-ALL cases harbored subclonal IGH or IGL rearrangements whereas in BL IG-MYC-r is considered a founder oncogenic event and is therefore clonal. We also identified recurrent i(1)(q10). While gains of 1q are the most common abnormality in BL, i(1)(q10) has not been previously reported in that disease.^29-32 Equally, we identified only two non-IG-MYC-r patients with i(1)(q10) within the Leukaemia Research Cytogenetics Group (LRCG) database of >12,000 cases of BCP-ALL and while a recently published cohort of ALL had two cases with i(1)(q10) among 18 identified as harboring a BCL2/BCL6/MYC-r, only two other cases with i(1)(q10) were seen among the remaining 1,970 cases.¹⁴ Interestingly, isolated cases of immunophenotypically immature BCP-ALL with IG-MYC-r and i(1)(q10) have been reported previously.^11,16 While these data support a specific role for i(1)(q10) in BCP-ALL with IG-MYC-r with or without IGBCL2-r, the biological implications of this specific finding are unknown. Our mutational analysis of cases confirmed recurrent KRAS mutations within this group^14,16 and identified additional recurrent mutations in the BCP-ALL-related genes IKZF1 and ASMTL.

Targeted sequencing of the IGH locus identified breakpoints predominantly within the V(D)J gene segments. This contrasts with germinal center-derived malignancies, including BL, in which just 9-19% of cases carry a V(D)J gene segment breakpoint with rearrangements instead predominantly affecting the constant/switch regions.^49-53 While our DNA methylation/transcriptome analyses were able to cluster cases with their underlying cytogenetic aberration (KMT2A-r or IGH-DUX4) when present, other IG-MYC-r cases with V(D)J gene segment breakpoints formed a separate cluster distinct from BL. In just two cases we identified IGH constant region breakpoints, co-occurring with the only examples in our study of typical BL-related mutations in ID3/TCF3, DDX3X, SMARCA4 and CCND3.^37-39 Furthermore, one of these two cases also had an IG-BCL2-r. DNA methylation analysis clustered all three cases displaying either IGH constant region breakpoints and/or IG-BCL2-r with BL cases/cell lines, in keeping with the recently described occurrence of BCL2-r in adult molecular BL.⁴⁰ However, these findings contrast with a subgroup of BCL2/BCL6/MYC rearranged ALL which were recently described to cluster together by RNA-sequencing analysis.¹⁴ Whether this relates to alternative experimental approaches/comparators, immunophenotypic inclusion criteria or in fact further demonstrates heterogeneity within this group of patients remains unclear. However, the recent publication of the clinical outcome of this subgroup further supports the dismal prognosis associated with BCL2/BCL6-r ALL.⁵⁴ Whether BCL2/BCL6-r identifies a different disease from the majority of cases of IG-MYC-r BCP-ALL remains to be determined but neither published data nor the data presented here support the treatment of these patients with standard BCP-ALL therapy (Figure 7C).

Given the inconsistent treatment protocols and poor follow-up associated with patients not enrolled on a clinical trial, we sought to define patients suitable for ALL therapy for whom inclusion in prospective clinical trials would offer standardized risk-stratified, response-adapted therapy with analysis of long-term follow-up data. The inclusion of cases with BCP-ALL and an ALL-specific cytogenetic abnormality is, we believe, uncontentious. Nonetheless, the presence of a MYC-r in these cases is bound to generate questions and therefore clinical trial inclusion/exclusion criteria need to be definitive. More challenging are cases in which IG-MYC-r is the only defining cytogenetic abnormality. Historically, these patients’ outcome has been very poor, but the very few data that exist predominantly predate minimal residual disease analysis and response adaptation. With small numbers of patients in both the group remaining on an ALL trial and those removed from trial to receive therapy for which the details are not available, we did not observe a significant difference in EFS or OS. While there was a non-significant trend towards improved outcome for those removed from trial protocol therapy, the impact of heterogeneous clinician-determined therapies was unclear. These therapies are likely to have included ALL or BL protocols but also hybrid treatments delivering intensified early “BL-type” therapy with subsequent addition of ALL maintenance chemotherapy.^17,18 These limited data underscore the need for a prospective analysis of outcome using a standardized protocol. The rarity of the cases makes it unfeasible to conduct a dedicated trial solely for these patients. Currently they are being lost to investigation and receive therapy of unproven efficacy resulting in outcomes significantly inferior to those of standard ALL or BL. We believe that the underlying biology of these IG-MYC-r patients indicate that they are part of the ALL spectrum. To improve our understanding of this BCP-ALL subtype and patients’ outcome, we strongly suggest that they be included in clinical trials able to prospectively measure early treatment response, supporting informed decision-making on whether to continue a standard treatment strategy or not. Those with very poor response to initial therapy may be considered for alternative therapies, similar to the approach to acute leukemia of ambiguous lineage being taken by some study groups.^55,56

Figure 7.Swimmer plots displaying event-free survival and overall survival. (A) Children and (B) adults with IG-MYC-rearrangements but no other acute lymphoblastic leukemia-specific cytogenetic abnormality. Suggested approach for the assignment of patients to a treatment strategy (C). + For patients without data on relapse or second tumor, death was assumed to be the first event. ALL: acute lymphoblastic leukemia; BCP: B-cell precursor; CCR: continuous complete remission; sIG: surface immunoglobulin.

Within that setting, use of disease-agnostic approaches such as antibody or advanced cellular therapies, may be logical.

In summary, we have demonstrated that surface immunoglobulin-negative BCP-ALL with IG-MYC-r are a high-risk subgroup of BCP-ALL. Optimal treatment can only be determined by removing barriers to treating patients with uniform therapy and collection of detailed longterm clinical follow-up data. We believe that clinical trial protocols should not exclude these patients, who should in fact be given access to contemporary, risk-stratified, response-adapted therapy.

Footnotes

• Received January 6, 2022
• Accepted March 31, 2022

Correspondence

*SB and LJR contributed equally to this work.

°current address University Medical Center Utrecht, Utrecht, the Netherlands

L.J. Russell lisa.russell@newcastle.ac.uk

S. Bomken s.n.bomken@newcastle.ac.uk

Disclosures

No conflicts of interest to disclose.

Contributions

SB, LJR, FWvD and AVM conceived and designed the study; SB, YD, AA, YB, JB, BB, GAAB, GC, GG, HAdG, CH, LLN, MP, AP, ESe, ESo, SS, VHJvdV, VR, SPH, CJH, FWvD, MLL, JV, AVM and LJR provided clinical data and contributed patient samples; SB, ECS, AM, MZ, KTMF, MB, HL, LJ, NK, EW, CA, CMB, BAW, AVM and LJR performed experiments and analyzed experimental data; AE and YD collected and analyzed outcome data; SB, LJR and AVM wrote the manuscript with additional contributions from MLL, JM and JV; all coauthors contributed to manuscript review and editing and approved the final manuscript for submission.

Data-sharing statement

All previously publicly available datasets used are referred to in the Methods section and in Online Supplementary Table S3. Methylation data are available from the GEO repository under accession number GSE174248. RNA, whole exome and targeted sequencing data are available from the EGA repository under EGA study ID EGAS00001005111.

Funding

We would like to thank those who provided financial support. Kay Kendall Leukaemia Fund (KKL515), Wellcome Trust Institutional Strategic Support Fund (204787/Z/16/Z), The Newcastle upon Tyne Hospitals NHS charity and KidsCan Children’s Cancer Research and The Little Princess Trust (to LJR). LJR held a Newcastle University Research Fellowship and A.M. a Newcastle University Faculty Fellowship. SB was supported by The Sir Bobby Robson Foundation and an MRC Clinician Scientist Fellowship MR/S021590/1. COG funding was available through U10 CA98413 and U10 CA180899 (COG Statistics and Data Center grants), U24 CA114766 and U24-CA196173 (COG Specimen Banking). AVM and CJH were supported by Blood Cancer UK (15036) and Cancer Research UK (A21019, Moorman and Fielding). LLN was partially supported by MFAG 2009 – AIRC (Italian Association for Cancer Research) (Lo Nigro). VR was supported by a Bloodwise (formerly Leukaemia and Lymphoma Research) Senior Bennett Fellowship (#12005), the JGW Patterson Foundation and the Little Princess Trust. MZ was supported by North East Promenaders. GC was supported by the Italian Association for Cancer Research (Grant n.IG17593) and Comitato Maria Letizia Verga.

Acknowledgments

We would like to thank patients and their families for donating their samples for research and each of the institutional/national BioBanks for providing those samples for this study, including the Blood Cancer UK Childhood Leukaemia Cell Bank. We thank Adele Fielding for help in identifying UK patients.