AUTHORS
Luiz Fernando Bazzo Catto, MD, PhD, Fernanda Gutierrez-Rodrigues, PhD2, Matthew Gianferante3, Alexandra Satty, MD4, Maryam Hussain, MPH5, Katie Mills, DNP6, Luiz Darrigo7, Marília Catto, MD8, Xing Fan, MD9, Hyun Sung Park, PhF10, Flavia Donaires11, Neal Stuart Young, MD12, Sharon A Savage, MD13, Neelam Giri, MBBS, MD14, Rodrigo T. Calado, MD, PhD15, Jeffrey Michael Lipton, MD, PhD16, Adrianna Vlachos, MD17, Cynthia Dunbar, MD18, David J Young, MD, PhD19
AFFILIATIONS
1 National Heart, Lung, and Blood Institute, Bethesda, Maryland, United States
2 National Heart, Lung, and Blood Institute, Bethesda, Maryland, United States
3 National Cancer Institute, Bethesda, Maryland, United States
4 The Feinstein Institutes for Medical Research, Manhasset, New York, United States
5 The Feinstein Institutes for Medical Research, Manhasset, New York, United States
6 Johns Hopkins All Children's Hospital, St. Petersburg, Florida, United States
7 Universidade de São Paulo, Ribeirao Preto, Brazil
8 University of São Paulo, Ribeirao Preto, Brazil
9 National Heart, Lung, and Blood Institute, Bethesda, Maryland, United States
10 National Heart, Lung, and Blood Institute, Bethesda, Maryland, United States
11 Universidade de São Paulo, Ribeirao Preto, Brazil
12 National Heart, Lung, and Blood Institute, Bethesda, Maryland, United States
13 National Cancer Institute, Bethesda, Maryland, United States
14 National Cancer Institute, Bethesda, Maryland, United States
15 Universidade de São Paulo, Ribeirao Preto, Brazil
16 Cohen Children’s Medical Center, New York, New York, United States
17 Cohen Children’s Medical Center, New York, New York, United States
18 National Heart, Lung, and Blood Institute, Bethesda, Maryland, United States
19 National Heart, Lung, and Blood Institute, Bethesda, Maryland, United States
ABSTRACT
Genetic testing for pathogenic germline variant (PGV) and cancer-associated somatic mutations is increasingly used in the molecular diagnosis of inherited bone marrow failure syndromes (IBMFS). Diamond-Blackfan Anemia Syndrome (DBAS) is an IBMFS characterized by hypoplastic anemia, congenital anomalies and cancer predisposition caused by PGV in ~30 small (RPS) or large (RPL) ribosomal subunit genes. Sequencing-based screening identifies pathogenic single-nucleotide variants (SNV), indels and copy number variations (CNV) in 70-80% of cases. The clinical utility of somatic testing for DBAS is uncertain as recurrent clonal hematopoiesis (CH) has not been identified, and myeloid neoplasia (MN) incidence is lower in DBAS than other IBMFS. We investigated whether an integrated targeted panel for germline, somatic, and CNV improves the molecular diagnostic rate and clinical monitoring for DBAS.
We developed a comprehensive error-corrected targeted sequencing panel (0.5% limit of detection) covering 181 genes implicated in DBA, MN, erythropoiesis or ribosome biology. We retrospectively sequenced DNA from granulocytes (NHLBI) or total leukocytes (NCI, USP, DBAR) from 212 DBA patients evaluated in 4 centers in the USA and Brazil. Genetic findings were correlated with clinical and laboratory data.
Median age was 11 years (range 0–57), 45% were female, and 173/212 (81%) had a previously known molecular diagnosis, most commonly involving RPS19 (37%), RPS26 (14%), and RPL5 (11%), most PGVs were SNV or indels (73%), with few splicing variants (17%) or CNVs (10%). Of the 39 patients without a molecular diagnosis found by prior genetic testing, our panel identified novel large deletions of RPS17, RPS19 and RPS26 in 6/39 patients (15%). Splicing and SNV PGVs in RPS17, RPS26, RPS19, RPS10, RPL26, and RPL27 were found in 13/39 previously unknown cases, indicating that method and calling discrepancies underestimate the molecular diagnostic rate in DBA.
Two patients with previously unidentified molecular defects harbored PGVs with “somatic-range” VAFs: RPL26 c.310-16_353del (VAF 30%) and RPS19 c.94TG>T (VAF 30%); 1 became treatment independent, and 1 was corticosteroid-refractory, respectively. For both, SNP-array confirmed copy-neutral loss of heterozygosity (CN-LOH). Additionally, 3 patients with previously-known PGVs were subsequently found to be at “somatic” ranges due to CN-LOH: 2 were steroid-refractory, transfusion-dependent (RPS19, VAF 31% and 39%, the latter was eltrombopag-responsive), 1 was corticosteroid-dependent (RPL35A, VAF 37%).
CH was detected in 27 (13%) patients (median age: 31 years vs. 10 years without CH, p=0.007). TET2- and DNMT3A-CH were detected in 4/27 (15%) patients (VAF 0.7–4.9%), all >39-years-old. PPM1D-CH was found in 1/27 patients (4%, VAF 1.8%), and TP53-CH was found in 5/27 patients (18%, VAF 0.6–3.1%); all were >25-years-old, transfusion-dependent, and MN-free at last follow-up. Notably, we identified a distinct CH signature characterized by enrichment of STAG2, ASLX1, and SETBP1 mutations in 4 patients, all with GATA1 PGVs known to generate a short isoform (GATA1s) that primarily affects erythropoiesis. Two GATA1s siblings carried multiple STAG2 mutations (>6 or >4, with VAFs ranging from 0.5–9.3% and 0.7–3.1%, respectively) co-occurring with ASXL1 (VAF 5.5%) or DNMT3A mutations (VAF 0.7%); only 1 had MN at time of sequencing. STAG2 (VAF 7.4%), ASXL1 (VAF 0.9–1.2%) and SETBP1 mutations (VAF 0.5%) were detected in 2 other GATA1s siblings who had MN at last follow-up. Further analysis is required to determine if these mutations were pre-malignant and associated with GATA1s progression to MN.
Unlike other IBMFS, CNV analysis yielded a molecular diagnosis in 23/212 (11%) DBAS patients. Inclusion of variants at VAFs >25% in the analysis allowed PGV detection in the presence of CN-LOH for 5/212 patients (2%). For RPS-/RPL-mutated DBAS, the clinical utility of somatic testing remains uncertain, as no patients developed MN in our cohort. In contrast, GATA1s PGV is associated with a STAG2, SETBP1, and ASXL1 CH signature, suggesting distinct pathophysiology with potential MN surveillance implications: 3/4 (75%) patients had MN diagnosis at last follow-up.
Integrated germline, somatic and CNV analysis improved molecular diagnostic yield in DBAS to 90% and may be particularly useful for identification of patients with GATA1s-PGVs for risk stratification and intensified surveillance.
DISCLOSURES
Luiz Fernando Bazzo Catto, MD, PhD: No relevant conflicts of interest to declare.