The genetic tumor background is an important determinant for heterogeneous MYCN-amplified neuroblastoma

Abstract

Amplification of MYCN is the signature genetic aberration of 20-25% of neuroblastoma and a stratifying marker associated with aggressive tumor behavior. The detection of heterogeneous MYCN amplification (hetMNA) poses a diagnostic dilemma due to the uncertainty of its relevance to tumor behavior. Here, we aimed to shed light on the genomic background which permits hetMNA in neuroblastoma and tied the occurrence to other stratifying markers and disease outcome. We performed SNP analysis using Affymetrix Cytoscan HD arrays on 63 samples including constitutional DNA, tumor, bone marrow and relapse samples of 26 patients with confirmed hetMNA by MYCN-FISH. Tumors of patients ≤18m were mostly aneuploid with numeric chromosomal aberrations (NCAs), presented a prominent MNA subclone and carried none or a few segmental chromosomal aberrations (SCAs). In older patients, tumors were mostly di- or tetraploid, contained a lower number of MNA cells and displayed a multitude of SCAs including concomitant 11q deletions. These patients often suffered disease progression, tumor dissemination and relapse. Restricted to aneuploid tumors, we detected chromosomes with uniparental di- or trisomy (UPD/UPT) in almost every sample. UPD11 was exclusive to tumors of younger patients whereas older patients featured UPD14. In this study, the MNA subclone appears to be constraint by the tumor environment and thus less relevant for tumor behavior in aggressive tumors with a high genomic instability and many segmental aberrations. A more benign tumor background and lower tumor stage may favor an outgrowth of the MNA clone but tumors generally responded better to treatment.

Keywords: MYCN amplification; intratumoral heterogeneity; neuroblastoma; uniparental disomy.

Figure 1

Clinical data associated with hetMNA in sample cohort. (a) Assigned stages of hetMNA tumors at diagnosis. (b) Age at diagnosis of the patients associated with clinical stage of the hetMNA tumors. The dashed line marks the 18‐month cutoff. Stages 2A and 2B were combined to stage 2 in subfigures a and b. (c) Localizations of primary tumors at diagnosis. The dashed line separates NB tumors with unknown location from those with a known primary location.

Figure 2

Summary of segmental chromosomal aberrations (>3MB) detected by SNP array analysis in hetMNA neuroblastoma samples. Samples are sorted according to patient age at diagnosis. The red dashed line marks the 18‐month age cutoff. Horizontal lines depict larger chromosomal aberrations and vertical lines amplifications or smaller deletions. Coloring of the boxes as described in the figure legend is based on the copy number state of the segments as determined by visual inspection of the SNP array data with the ChAS software. Presence of numerical aberrations in each sample is indicated by the fill of the circles to the right. Sequential breaks which result in the same type of aberration as classified in the legend but show different changes in copy number are indiscernible in the illustration. Due to the limited space, amplicons in close proximity cannot be visualized as separate entities (e.g., Pat.# 18, chromosome 12). For a detailed list of breaks and aberrations see Supporting Information Table 2. BM, bone marrow sample; cyto+, DNA was extracted from BM cytospin; MACS+, DNA from GD2‐positive cell fraction isolated by MACS from BM samples; TU, tumor sample.

Figure 3

SCAs, amplicons and UPDs. Number of segmental chromosomal aberrations (a) and breakpoints (b) in 26 hetMNA tumors was significantly different between the two patient age groups. Percentages of patients with indicated chromosomes affected by segmental aberrations (c) and gene amplifications (d). Incidences of SCAs in any tumor cell‐containing sample were considered once per patient. (e) The stacked column chart shows the distribution of uniparental whole chromosome isomies in the patient cohort. Presence of UPDs or UPTs were counted as an event regardless of penetrance in the tumor. Patients were grouped based on age at diagnosis: ≤18 months and >18 months. (f) UPD occurrence in 25 hetMNA tumors with UPD data grouped by tumor ploidy: di‐/tetraploid (6) and aneuploid (19). (g) UPD occurrence in 120 aneuploid NB tumors in the CCRI database grouped by MYCN status: non‐MNA (87), hetMNA (19) and MNA (14). (h) occurrence of UPD11 among the tumors of the 62 patients that carried UPDs: non‐MNA (41), hetMNA (14) and MNA (7). Numbers in round brackets represent patient counts.

Figure 4

Hypothetical model of proposed tumor evolution of hetMNA tumors. This model is based on the comparison of clinical and biological parameters of hetMNA NB tumors (25) included in this study and homMNA tumors (47) with evaluable SNP array data collected at the CCRI. Green cells depict favorable MYCN non‐amplified tumor cells in tumors of ≤18m patients whereas blue cells represent non‐MNA NB cells with a highly rearranged genome and a multitude of SCAs in tumors of older patients. Red illustrates MNA cells. Arrows depict a possible transition of aneuploid hetMNA tumors into aneuploid homMNA tumors which can be suspected based on similarities in the genetic background. The fact that most aneuploid homMNA tumors were found in >18m patients would suggest a time‐dependent clonal takeover of homMNA cells.