Introduction: Diamond-Blackfan anemia (DBA) is a rare inherited bone marrow failure syndrome (IBMFS) characterized by erythroid hypoplasia. It is associated with a number of congenital anomalies and a high risk of developing specific cancers. DBA is caused by germline mutations or deletions in genes affecting ribosomal biogenesis and function, with autosomal dominant or X-linked recessive patterns of inheritance. The most commonly mutated gene is RPS19, seen in approximately 25% of patients. About 45% of DBA families have no known disease-causing pathogenic variant.

Methods: Affected and unaffected individuals from families with DBA were ascertained through the IRB-approved NCI IBMFS retrospective/prospective cohort study (ClinicalTrials.gov Identifier: NCT00027274). Study participants completed detailed family and medical history questionnaires, medical records were reviewed, and a subset of families underwent clinical evaluations at the NIH Clinical Center. DBA patients enrolled prior to 2014 underwent routine clinical mutation testing for the established DBA genes; beginning in 2014, DBA patient samples (buccal and blood DNA) were evaluated by whole exome sequencing (WES) for mutation identification. We incorporated WES with deletion analyses and copy number variant (CNV) assessment to uncover the genetic changes causative of DBA. Deletion analyses performed included SNP genotyping and array comparative genomic hybridization. Functional effects of the genetic variants were proven by pre-rRNA processing defect analysis by Northern blot. Controls for functional studies were healthy mutation-negative individuals from the IBMFS study.

Results: Genetic testing information was available in 61 of the 87 families with DBA enrolled in the IBMFS study. Thirty-five of the 61 families did not have a known genetic cause at enrollment. Our combined approach of WES, deletion and CNV analyses identified the causative pathogenic variant in 18 of the 35 (51%) uncharacterized DBA families. We discovered pathogenic variants in two previously undescribed genes in two DBA families. One family had a nonsynonymous variant (p.K77N) in RPL35; the second family had a nonsynonymous variant (p. L51S) in RPL18. Both of these variants result in characteristic pre-rRNA processing defects. Our analyses also uncovered germline mosaic deletions in known DBA genes in both buccal and blood cells of two patients from two different families. One was a 1.8 Mb mosaic deletion in chromosome 15 including RPS17; the other was a large 2.5 Mb mosaic deletion on chromosome 3 including RPL35A. In addition to these findings, we found variants in previously known DBA-associated ribosomal genes in 14 of the 35 families.

We further evaluated the genomic characteristics of the entire DBA cohort. Pathogenic variants in ribosomal DBA genes were found in a total of 44 of the 61 families (72%) on whom genetic testing information and/or biospecimens were available. RPS19 was the most frequently mutated gene and accounted for 36% of families, followed by RPL35A and RPS26, accounting for 14% and 11% each, respectively. Notably, 30% of the variation in disease-causing genes in our cohort was due to a single copy or mosaic gene deletion. We had complete parental testing and inheritance information on 23 (52%) of the 44 families whose gene was identified. Ten of the 23 (43%) had an inherited mutation and 13 (57%) had a de novo change in the causative gene (both parents were negative for the affected child's disease-associated mutation). At this time, 17 of 61 families tested (28%) do not have a characterized disease-associated mutation.

Conclusion: This efficient comprehensive genomic approach was the basis for our discovery of two novel causes of DBA, characterization of ribosomal gene deletions not previously described to be disease-associated, and of DBA-associated germline mosaicism. We identified the disease-associated mutations in 51% (18 of 35) of our families without a known genetic cause of DBA. A total of 74% (44 of 61) of our families are now genetically characterized. Our comprehensive approach appears to provide more genomic information than other methods since pathogenic variants of DBA genes have been reported previously in about 55% of DBA patients.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.