ALBERT

Description: Abstract 2369 Diamond-Blackfan anaemia (DBA) is a rare autosomal dominant disorder associated with inactivating mutations in ribosomal protein (RP) genes, causing defects in erythroid progenitor and precursor cell development. Many cases are due to de novo mutations and in family cases there is often clinical heterogeneity due to variable penetrance. Mutations in RPS19 account for 25% of all DBA cases and single nucleotide variations (SNV), indels and allele-loss deletions have been found in 11 other RP genes in a further ∼50% of patients. Around 25% of patients with DBA have no identifiable mutations. Given that (with the exception of 2 cases with GATA1 mutations) all mutations in DBA characterised so far affect RP genes, it is likely that mutations in one of the 80 RP genes will be eventually identified in a significant proportion of the patients. Current screening methods are primarily based on Sanger sequencing on a per-exon/per-gene basis, with the associated time, labour and cost restrictions. We therefore aimed to evaluate high-throughput sequencing technology, including a bespoke target enrichment platform, to screen all 80 known RP genes to facilitate rapid, cost-effective identification of DBA associated mutations. DNA was extracted from peripheral blood samples that had been referred to Imperial Molecular Pathology for DBA screening from 10 individuals, including 3 family pairs: affected mother and daughter; 2 affected siblings; and another sibling pair, one of whom was unaffected/low-penetrance (no defining clinical symptoms, except for high adenine deaminase). Only one patient had a known mutation (RPS19 c.280C〉T) and was included as a control. Agilent SureSelect XP was used for the target enrichment, which employed a custom designed tiled-RNA bait hybridisation solution to capture the target genes, including non-masked intronic regions and 500bp of flanking sequence. The DNA was sheared using a Covaris e220, QC was performed via QIAxcel capillary electrophoresis and the hybridisation was carried out at 65°C for 48h. Individual libraries were quantified using qPCR against the supplied standard curve and pooled proportionally. The sequencing was performed on an Illumina MiSeq, using 150bp paired-end reads and multiplexed using the supplied ScriptSeq barcodes. The sequencing reads were aligned to the build 37 reference genome using BWA software, and the variant calls made using GATK. Annovar was used for functional annotations of the variants. Protein truncating mutations were found in RP genes in 7 of the 10 samples, including the positive control and 6 of the 8 clinically confirmed DBA patient samples., All mutations were in RP genes previous described as being involved in DBA, although 3 affected novel codons: RPL5 c.G244T (stop-gain SNV; novel; mother-daughter pair); RPL5 c.166_169delACAA (frameshift); RPS10 c.C337T (stop-gain SNV); RPL11 c.472–473delAA (frameshift; novel); RPS26 c.212–213insA (frameshift; novel). Validation was by Sanger sequencing and further confirmation testing will include unaffected family members. The remaining 2 DBA patients, a brother-sister pair, showed no definable mutations in the captured regions and neither did the unaffected/low-penetrance sibling of the RPS10 patient. In summary, a rapid and cost effective methodology for screening genetic lesions associated with the causation of DBA is warranted, especially given the magnitude of attaining global coverage by conventional techniques. Whole-gene enrichment followed by multiplexed runs on a bench-top class high-throughput sequencing platform is arguably the approach of choice; although as the cost of exome and even genome sequencing continues to fall, these may well become realistic options in the coming few years. This work is ongoing, with a second group of 10 samples already sequenced and undergoing analysis, and bioinformatic refinements, especially for the detection of larger deletions, may yet yield results for the two undetected samples. These preliminary results suggest that high throughput sequencing technology with a bespoke target enrichment platform for RP genes is a feasible, efficient and relatively rapid diagnostic tool for detection of causative mutations DBA. Disclosures: No relevant conflicts of interest to declare.

Description: The thalassemias are inherited disorders classified genetically into α, β, γ, δβ, δ and εγδβ varieties according to the type of globin(s) that are underproduced. At the molecular level, the εγδβ thalassemias fall into two categories; Group I removes all, or a greater part, of the β globin gene cluster which is embedded in an array of olfactory receptor genes on chromosome 11p15. Group II removes extensive upstream regions leaving the β globin gene itself intact despite which, its expression is silenced due to inactivation of the upstream locus control region (β LCR). Recently, two novel deletions causing εγδβ thalassemia have been reported; a 153 kb deletion removing the entire β globin cluster in a Chilean family (Game, L., et al., Br J Haematol2003, 123:154–9) and an upstream deletion of 112 kb in a Dutch family (Dutch III) (Harteveld, C.L., et al., Br J Haematol2003,122: 855–8). We describe here the characterization of another three novel εγδβ thalassemia deletions, in three English families, named English II, III and IV, to differentiate them from the previously reported English (I) deletion (Curtin, P., et al., J Clin Invest1985, 76: 1554–8). Deletion English II removed 98 kb extending 90 kb upstream of the ε gene to 8 kb upstream of the Gγ gene, and included 4 upstream olfactory receptor (HOR) genes. Deletion English III removed 114 kb extending 60 kb upstream of the ε gene to 9 kb downstream of the β globin gene, thus including the entire β globin gene cluster as well as two upstream HOR genes. English IV is the largest deletion (439 kb) reported so far; starting 326 kb upstream of the ε gene to 70 kb downstream of the β gene and included 13 upstream, and 3 downstream, HOR genes plus the intervening β globin gene cluster. Breakpoints of all the 3 deletions occurred within regions of L1 or Alu repeat elements and contained short regions of direct homology between the flanking sequences, a feature that is likely to have contributed to the illegitimate recombinations. Deletions English II and III appear to be de novo while English IV is not. The proband for the English IV deletion had neonatal hemolytic anemia and required blood transfusions while 3 other family members who were heterozygous for the same deletion, had uneventful post-natal periods. The English III proband also required a blood transfusion soon after birth while the English II proband did not. Although in later life, heterozygotes for εγδβ thalassemia are transfusion-independent, and have a blood picture typical of β thalassemia trait but with normal Hb A2 levels, our data suggest that heterozygotes for εγδβ thalassemias have more severe microcytosis and hypochromia than β thalassemia carriers. To date, a total of 15 deletions causing εγδβ thalassemia have been described - five upstream deletions (Group II) associated with intact β globin genes and ten (Group I) that include the entire β globin gene cluster. These deletions are all unique and illustrate the heterogeneity of the εγδβ thalassemias.