High resolution SNP array profiling identifies variability in retinoblastoma genome stability

Abstract

Both hereditary and nonhereditary retinoblastoma (Rb) are commonly initiated by loss of both copies of the retinoblastoma tumor suppressor gene (RB1), while additional genomic changes are required for tumor initiation and progression. Our aim was to determine whether there is genomic heterogeneity between different clinical Rb subtypes. Therefore, 21 Rb tumors from 11 hereditary patients and 10 nonhereditary Rb patients were analyzed using high-resolution single nucleotide polymorphism (SNP) arrays and gene losses and gains were validated with Multiplex Ligation-dependent Probe Amplification. In these tumors only a few focal aberrations were detected. The most frequent was a focal gain on chromosome 2p24.3, the minimal region of gain encompassing the oncogene MYCN. The genes BAZ1A, OTX2, FUT8, and AKT1 were detected in four focal regions on chromosome 14 in one nonhereditary Rb. There was a large difference in number of copy number aberrations between tumors. A subset of nonhereditary Rbs turned out to be the most genomic unstable, while especially very young patients with hereditary Rb display stable genomes. Established Rb copy number aberrations, including gain of chromosome arm 1q and loss of chromosome arm 16q, turned out to be preferentially associated with the nonhereditary Rbs with later age of diagnosis. In contrast, copy number neutral loss of heterozygosity was detected mainly on chromosome 13, where RB1 resides, irrespective of hereditary status or age. Focal amplifications and deletions and copy number neutral loss of heterozygosity besides chromosome 13 appear to be rare events in retinoblastoma.

RB1mutation spectrum in a comprehensive nationwide cohort of retinoblastoma patients

“…With the MYCN amplified tumours excluded, we detected 96% (96/100) of expected RB1 mutations in tumours. In the majority of tumours, the second hit was LOH (34/64%), in line with previous reports 17 26 34 35…”

Overview of recurrent chromosomal losses in retinoblastoma detected by low coverage next generation sequencing

“…Other gains are in Chr1, 6, 7, 9, 11, 12, 16, 17, 19, and 22 although they were of lower magnitude and more discontinuous when compared with the losses observed in Chr13 or ChrX. Since we know that losses detected with this pooling approach reveal recurrent losses but not recurrent gains among the pooled tumors, we then focused the analysis on recurrent losses in all other chromosomes finding indeed, recurrent losses previously reported by other authors in all chromosomes (14,37,38). Patterns of similar losses between Rb pools 1 and 2 can be appreciated across the chromosomes, suggesting that recurrent losses are not random, and even though the patterns between Rb pools can be discerned, they tend to be larger in pool 1, and although scattered more gained regions are also present in pool 1.…”

miRNome landscape analysis reveals a 30 miRNA core in retinoblastoma

“…With a map that we had previously generated using NGS which demonstrated regions of recurrent losses in Rb [29], we identified in the 995 undetectable miRNAs, those located at loci recurrently lost in Rb by chromosome. In total, we found 144 miRNAs located in areas recurrently lost in Rb [37, 38] the complete list is shown in Additional file 3. Data plotted in Fig.…”