We have analyzed RNA from retinoblastoma patients and unaffected carriers with various RB1 gene mutations to determine the patterns of missplicing and associations with phenotypic expression. Most sequence alterations in or in the neighborhood of conserved splice signals that we tested resulted in simple exon skipping (15 mutations) or intron inclusion (new acceptor AG-sites, four mutations) as expected. Two mutations resulted in skipping of a neighboring exon (exon 11), a complex pattern indicating competition for correct lariat formation. We observed no activation of a cryptic splice site but found that a recurrent missense mutation in exon 7 creates a new splice site (two families). RT-PCR analysis enabled us to confirm the presence and to characterize the transcriptional consequences of gross insertions and deletions in the RB1 gene in six patients, including two patients with mutational mosaicism. We also used RT-PCR analysis to search for unknown mutations in 15 patients and identified three oncogenic point mutations deep in introns. Two of these mutations are recurrent thus indicating that, despite the vast extent of the introns of the RB1 gene, few bases are effective targets for oncogenic mutations. When analyzing associations between phenotypic expression (16 families) and mutational consequences we observed no link to the presence or absence of a premature termination codon in the mutant transcript. However, the location of a mutation relative to the splice sequence has a strong and consistent influence on phenotypic expression. Hum Mutat 29(4), 475–484, 2008.
Elevated levels of cell‐free DNA (cfDNA) are frequently observed in tumor patients. Activating mutations in exon 4 (R183) and exon 5 (Q209) of GNAQ and GNA11 are almost exclusively found in uveal melanoma, thus providing a highly specific marker for the presence of circulating tumor DNA (ctDNA). To establish a reliable, noninvasive assay that might allow early detection and monitoring of metastatic disease, we determined the proportion of GNAQ or GNA11 mutant reads in cfDNA of uveal melanoma patients by ultradeep sequencing. Cell‐free DNA from 28 uveal melanoma patients with metastases or extraocular growth was isolated and quantified by real‐time polymerase chain reaction (PCR) (7–1550 ng DNA/mL plasma). GNAQ and GNA11 regions of interest were amplified in 22 of 28 patients and ultradeep sequencing of amplicons was performed to detect even low proportions of mutant reads. We detected Q209 mutations (2–38% mutant reads) in either GNAQ or GNA11 in the plasma of 9 of 22 metastasized patients. No correlation between the proportion of mutant reads and the concentration of cfDNA could be detected. Among the nine ctDNA‐positive patients, four had metastases in bone, whereas no metastases were detected in the 13 ctDNA‐negative patients at this location (P = 0.025). Furthermore, ctDNA‐positive patients tended to be younger at initial diagnosis and show larger metastases. The results show that ultradeep amplicon sequencing can be used to detect tumor DNA in plasma of metastasized uveal melanoma patients. It remains to be shown if this approach can be used for early detection of disseminated tumor disease.
Background: Uveal melanoma (UM) is a rare eye tumor. There are two classes of UM, which can be discriminated by the chromosome 3 status or global mRNA expression profile. Metastatic progression is predominantly originated from class II tumors or from tumors showing loss of an entire chromosome 3 (monosomy 3). We performed detailed EFS (embryonal Fyn-associated substrate) methylation analyses in UM, cultured uveal melanocytes and normal tissues, to explore the role of the differentially methylated EFS promoter region CpG island in tumor classification and metastatic progression. Methods: EFS methylation was determined by direct sequencing of PCR products from bisulfite-treated DNA or by sequence analysis of individual cloned PCR products. The results were associated with clinical features of tumors and tumor-related death of patients. Results: Analysis of 16 UM showed full methylation of the EFS CpG island in 8 (50%), no methylation in 5 (31%) and partial methylation in 3 (19%) tumors. Kaplan-Meier analysis revealed a higher risk of metastatic progression for tumors with EFS methylation (p = 0.02). This correlation was confirmed in an independent set of 24 randomly chosen tumors. Notably, only UM with EFS methylation gave rise to metastases. Real-time quantitative RT-PCR expression analysis revealed a significant inverse correlation of EFS mRNA expression with EFS methylation in UM. We further found that EFS methylation is tissue-specific with full methylation in peripheral blood cells, and no methylation in sperm, cultured primary fibroblasts and fetal muscle, kidney and brain. Adult brain samples, cultured melanocytes from the uveal tract, fetal liver and 3 of 4 buccal swab samples showed partial methylation. EFS methylation always affects both alleles in normal and tumor samples. Conclusions: Biallelic EFS methylation is likely to be the result of a site-directed methylation mechanism. Based on partial methylation as observed in cultured melanocytes we hypothesize that there might be methylated and unmethylated precursor cells located in the uveal tract. The EFS methylation of a UM may depend on which type of precursor cell the tumor originated from.
Several tumors, including uveal melanoma, show somatic mutations of GNAQ/GNA11. Circumscribed choroidal hemangioma is a benign tumor that becomes symptomatic in adulthood. In some patients, morphologic examination of biopsies is required for differential diagnosis between amelanotic choroidal melanoma and circumscribed choroidal hemangioma. Here, we report the results of GNAQ/GNA11 mutation analysis in samples from circumscribed choroidal hemangioma. Deep amplicon sequencing (Illumina MiSeq, San Diego, CA, USA) of positions R183 and Q209 of GNAQ and GNA11 in tissue samples from 33 patients with histologically diagnosed circumscribed choroidal hemangioma. All patients underwent biopsy or enucleation at our clinic between 2008 and 2018. To enable detection of variant alleles at low fractions, read depth exceeded 15,000-fold. DNA for genetic analysis was prepared from either snap-frozen (n = 22) or FFPE (n = 11) tissue samples. Samples from 28/33 patients (85%) showed a somatic missense mutation of GNAQ (c.626 A > G) predicted to result in p.Q209R. Variant allele fraction was variable (range 2.3% to 28%). Variants of GNAQ resulting in p.Q209 are characteristic for circumscribed choroidal hemangiomas. It appears that the GNAQ mutation spectrum in this tumor is narrow, possibly restricted to p.Q209R. Moreover, the spectrum is distinct from that of uveal melanoma, in which alterations resulting in p.Q209R are very rare.