Introduction

MYC rearrangements (MYC-R) constitute an integral defining feature in the diagnostic classification of mature aggressive B-cell lymphoma (BCL) [1,2,3]. Specifically, diffuse large B-cell lymphoma (DLBCL)/high-grade B-cell lymphoma (HGBCL) with MYC and BCL2 rearrangements (and/or BCL6 rearrangements) circumscribe a subset of higher-risk tumors. Current guidelines recommend investigating MYC-R with fluorescence in situ hybridization (FISH) [4]. In accordance with the diversity of rearrangement partners including immunoglobin (IG) and non-IG partners and the variability of breakpoints in the MYC locus, a break-apart (BAP) FISH probe is commonly utilized. Dual-color dual-fusion FISH probes (D-FISH) spanning MYC and IGH, IG-lambda (IGL) or IG-kappa (IGK) may also be used. We have previously demonstrated that in suspected HGBCL unbalanced rearrangements are identified in 11.9% of cases with abnormal results with the MYC BAP probe. In approximately 8.5% of cases, these cannot be reconciled with IGH/MYC D-FISH results and remain of ambiguous significance [5]. Forty-three percent of unbalanced cases also harbor a concurrent BCL2 rearrangement (unpublished data), thereby emphasizing the importance of informed interpretation of results for accurate diagnostic classification and therapeutic management. In this study, we sought to elucidate the significance of these unbalanced rearrangements with whole genome sequencing (WGS) analysis.

Materials and methods

Survey

An online survey comprising 6 questions was distributed to the International Cytogenetic community via email (Supplementary Methods) to delineate the scope of FISH strategies used to investigate MYC-R and the interpretation practices of unbalanced MYC BAP results across clinical laboratories.

FISH analysis

FISH analysis consisted of commercial MYC BAP and MYC/IGH D-FISH probe sets (Abbott Laboratories, Des Plaines, IL). The MYC BAP probe set included a red (R) and a green (G) probe which respectively hybridized 5′ and 3′ to the MYC gene, yielding a fusion (F) signal in the setting of an intact MYC locus. While typical MYC-R are indicated by balanced, separate red (R) and green (G) signals (RGF-type pattern), unbalanced patterns represent unbalanced or isolated R or isolated G signals. Cases with isolated R signals in the absence of isolated G signals (such as 1R1F) were referred to as RF-type patterns, and cases with isolated G signals in the absence of isolated R signals (such as 1G1F), were referred to as GF-type patterns (Fig. 1A).

Fig. 1: Representation of findings obtained with whole-genome sequencing and correlation with fluorescence in situ hybridization results.
figure 1

A Representation of a typical red (R)-green(G)-fusion(F)-type pattern, B Representation of an unbalanced RF-type pattern, C Representation of an unbalanced GF-type pattern, D Illustration of breakpoints identified at the MYC locus in our cohort with whole-genome sequencing (GRCh38 reference genome, created with Gviz and trackViewer R packages), E Circos plot illustrating MYC rearrangement partners identified in our cohort (created with circlize R package).

Full size image

Whole genome sequencing

WGS was performed with DNA extracted from formalin-fixed, paraffin-embedded (FFPE) sections using Qiagen AllPrep or DNA FFPE kits (Cats #80234, #56404). A modified Covaris fragmentation protocol designed to capture larger insert sizes was used [6]. Libraries were multiplexed on an Illumina NovaSeq S4. Mapping to the GRCh38 reference genome and structural variant calling were performed with BIMA 3.1.5/SVAtools pipeline [6]. FFPEseq mean and range of uniquely mapped fragments for the 14 libraries was 492 M (386M-685M). Tumor bridge coverage (average number of fragments (read-pairs) spanning a position in the genome), adjusted for library insert length and pipeline estimated tumor percentage, was 30.5× (14.2×–61.4×).

Clinical evolution

Baseline demographic characteristics, management approaches and response to treatment were extracted from medical chart review.

This study was approved by the Mayo Clinic Institutional Review Board (15-007359).

Results

Survey of interpretation practices of unbalanced MYC break-apart results

Fifty-four responses were obtained to the survey querying laboratory practices regarding FISH strategies and interpretation of unbalanced MYC BAP results. The survey participants were derived from ≥ 31 different institutions located in ≥4 different countries (23 responders did not provide information related to their work institution). Twenty-three of 54 laboratories (43%) only performed the MYC BAP probe and 30/54 (56%) laboratories performed the MYC BAP and IGH/MYC D-FISH probes upon initial investigation. One of 54 laboratories (2%) performed the MYC BAP, IGH/MYC, IGK/MYC and IGL/MYC D-FISH probe sets upon initial investigation. BCL2 and BCL6 rearrangements were sought upfront by 36/54 (67%) responders and 14/54 (26%) only queried these rearrangements in the event of a MYC-R. Thirty-six percent and 42%, respectively, reported interpreting RF- and GF-type patterns as a MYC-R, 58 and 56% reported RF- and GF-type patterns as equivocal, respectively, and 6 and 2% reported RF and GF-patterns as negative for a MYC-R, respectively (Supplementary Fig. 1). These data demonstrate significant variability in the interpretation of unbalanced MYC BAP results with most laboratories reporting an equivocal result.

Cohort description and associated FISH analysis results

Our cohort included 14 cases of DLBCL/HGBCL evaluated in our clinical FISH laboratory between 2019 and 2021 and selected sequentially in inverse chronological order of sampling (with preference given to internal cases for clinical correlation). Seven cases had an unbalanced MYC BAP result, including 5 cases with a RF-type pattern and 2 cases with a GF-type pattern on MYC BAP analysis in the absence of an IGH partner by D-FISH. Seven specimens with a typical RGF-type pattern with a known MYC-IGH (n = 1) or unknown (n = 6) partner were also included. These served as controls to verify the ability of the WGS methodology from FFPE sections used in this study to identify rearrangements and dissect the genomic architecture at the MYC locus.

Whole genome sequencing results

A MYC-R was confirmed in all control cases with a balanced pattern on BAP FISH (Table 1, Fig. 1). These involved previously reported rearrangement partners/loci (IGH (n = 1), IGL (n = 2), ZCCHC7 (n = 1), RFTN1 (n = 1), DMD (n = 1) and BCL11A (n = 1)). A structural variant (SV) involving the MYC region was also detected in all 5 cases with a RF-type pattern. In line with the higher number of R signal(s) observed on BAP FISH, a relative gain of genomic material 5′ of the MYC gene or relative loss of material 3′ of the MYC gene was detected with WGS in all these cases. Putative fusion events juxtaposing MYC to (1) an intergenic region upstream of TG, (2) TG, (3) IRF8, and (4) SPAG1 was detected in 4/5 cases, while case 5 involved a copy number (CN) gain of MYC. WGS also allowed to resolve the unbalanced FISH results for cases with a GF-type pattern and revealed SVs leading to a higher copy number (CN) of the 3′MYC BAP FISH probe-binding sequence in comparison with the 5′ region in both cases with a GF-type pattern, thereby also reconciling BAP FISH results. The first case with a GF-type pattern juxtaposed MYC with ACTB and the second involved a CN loss including MYC and the 5’ FISH probe-binding sequence. Of all cases with unbalanced FISH results, this case represented the only one in which WGS revealed a deletion involving MYC in our study cohort. MYC overexpression by immunohistochemistry ( ≥ 40%) was detected in 4/5 and 2/2 cases with RF- and GF-type patterns, respectively, suggesting that in 6/7 cases, the unbalanced MYC rearrangement was associated with increased MYC expression. In 7 cases with a typical pattern, overexpression of MYC was documented by IHC.

Table 1 Correlation of fluorescence in situ hybridization and whole-genome sequencing results at the MYC locus in 14 patients with DLBCL/HGBL.
Full size table

Clinical correlation

Clinical information was available for 3 cases with a RF-type pattern (Supplementary Table 2). While one case exhibited a favorable response to MR-CHOP chemotherapy, the other two cases remained refractory to therapy (R-CHOP and R-CHOP followed by R-ICE, polatuzumab+bendamustine respectively) and expired from disease. In the first case with a GF-type pattern, a favorable response to R-CHOP chemotherapy was exhibited; nonetheless, the patient expired from sepsis and multi-organ failure. In the second case, no systemic therapy was provided to the patient.

Discussion

The relevance of elucidating the implications of unbalanced MYC BAP results is underscored by the variability of interpretative practices across clinical laboratories highlighted by the survey we distributed to the cytogenetics community. Our study reveals the presence of true SV juxtaposing the MYC locus with a partner gene in most of these cases (4/5 with a RF-type pattern and ½ with a GF-type pattern). The remaining two cases (1/5 with a RF-type pattern and 1/2 with a GF-type pattern) also displayed a SV involving the MYC locus, yet these consisted of CN alterations and would not be considered as MYC-R per current DLBCL/HGBCL classification schemes [1,2,3]. In all cases, material 5′ of MYC for RF-type patterns and 3′ of MYC for GF-type patterns was present at a higher copy number state relative to 3′ and 5′ regions, thereby explaining the unbalanced FISH results.

MYC-R in DLBCL/HGBCL are thought to be acquired through aberrant activation-induced cytidine deaminase (AICDA)-mediated somatic hypermutation (SHM) and class-switch recombination (CSR) [7]. While AICDA mediates SHM and CSR at the IG loci, off-target mutagenic activity of this enzyme may occur in lymphoma-associated oncogenes such as MYC and result in oncogenic rearrangements. Breakpoints associated with MYC rearrangements exhibit significant variability within the MYC region. They may occur in a region designated as the “genic cluster” which encompasses a segment ~1.5 kb upstream of the transcription start site and the first exon and intron of MYC [8]. In our cohort, one case with a GF-type pattern exhibited a breakpoint within intron 1, a mechanism which has been suggested to result in aberrant MYC expression through dissociation of the natural promoter regions P1 and P2 and regulatory sites, resulting in transcription initiation from a cryptic promoter P3 in intron 3 [9]. Breakpoints located downstream of MYC may result in aberrant activation through the acquisition of MYC super-enhancers within a topologically-associated domain (TAD) comprising MYC, thus allowing for MYC activation through long-range loop interactions favored by a common CTCF binding site located 2 kb upstream of MYC [10]. Heterologous enhancers conferred by the rearrangement event may promote aberrant MYC expression in this setting [11]. In line with these considerations and in support of their potential for MYC activation, 4/5 RF-type cases involved breakpoints located downstream of MYC.

By study design, all cases with unbalanced FISH results and a confirmed MYC-R by WGS involved a non-IG partner gene. Consequently, the potential significance of these rearrangements must be taken into the context of the unclear prognostic relevance of non-IG MYC-R [12, 13]. An additional interpretative consideration of cases with unbalanced FISH results relates to the observation that these involved novel non-recurrent non-IG partner genes. Accordingly, in the absence of RNA expression data, whether these result in MYC overexpression remains difficult to ascertain, yet these were associated with MYC overexpression by IHC. Also, recently Collinge et al. [14] demonstrated that unbalanced patterns with loss of R or G signals were associated with mRNA expression levels comparable to HGBCL with MYC and BCL2 rearrangements exhibiting balanced MYC FISH results. Similarly, several of the BCL2 rearrangement in RF-type cases involved non-IG partners. In light of a paucity of data on atypical BCL2 rearrangement partners, their clinical significance remains ambiguous. Our study is further limited by a limited sample size which precluded a refined analysis of potential differential implications of different unbalanced patterns by FISH. A recent study suggested that while patterns with 5′MYC gains may adversely impact prognosis, a loss of 3′MYC may not portend similar deleterious implications [15].

In all, our study reveals that true genomic SV involving the MYC locus often underlie unbalanced MYC BAP FISH results. While these SV may result in a juxtaposition of MYC with a rearrangement partner, less frequently, they may also be associated with a CN alteration at the MYC locus. Accordingly, our results counsel caution in the interpretation of the significance of these unbalanced BAP MYC signal patterns. While larger studies are needed to validate our findings, an interpretation rendered as ‘likely positive’ for a rearrangement at the MYC gene locus may be most judicious. While the significance of an unbalanced pattern by MYC BAP analysis cannot be ascertained with BAP FISH testing alone and data remains currently limited, most cases have been shown to represent true rearrangements at the MYC gene region (Fig. 2). This may allow communication of the atypical nature of results while also providing added context reflecting the current body of evidence for the provider. Additional testing may assist in further clarifying whether a true rearrangement juxtaposing MYC with a gene partner and resulting in MYC overexpression is present.

Fig. 2: Proposed algorithm to guide the interpretation of unbalanced results obtained with the MYC break-apart (BAP) probe based on study results.
figure 2

Unbalanced patterns include RF (red-fusion)-type and GF (green-fusion)-type patterns. See text for further information. *It has previously been shown that evaluation for MYC rearrangements using only the MYC BAP is associated with a false negative (FN) rate of at least 4%. This FN rate can be reduced using IG/MYC D-FISH probes [16]. DF: probe dual color, dual-fusion probe set, IHC immunohistochemistry, RNAseq RNA sequencing, SV structural variant, WGS whole genome sequencing.

Full size image