Detection of Copy Number Alterations in DNA Repair/Repair-Related Genes for a Diversity of Hematologic Malignancies Identifies Disease-Specific Biomarkers of Potential Clinical Relevance.

Abstract 2531

DNA repair and repair-related genes encode molecules that function as critical guardians of the genome, both in terms of chromosome stability and single base integrity. Defects in such functions, both inherited and acquired, represent key factors that contribute to the etiology of cancer (e.g., RB1, TP53 and ATM). Copy number alterations (CNAs) in such genes have been described for one or more hematologic malignancies and disease-specific testing is routinely employed for the detection of these important diagnostic/prognostic markers. Additional DNA repair genes have been implicated as potential factors influencing the onset, progression, and/or chemo-resistance of disease. In the present study, we analyzed 155 DNA repair/repair-related genes for CNAs in 425 cases of hematologic malignancy using microarray analysis. Cases represented nine distinct clinical entities with at least 30 cases per entity. Disorders represented include myelodysplastic syndrome (MDS), chronic lymphocytic leukemia (CLL), myeloproliferative neoplasm (MPN), chronic myeloid leukemia (CML), B-cell and T-cell acute lymphoblastic leukemia (B-ALL, T-ALL), acute myeloid leukemia (AML), non-Hodgkin lymphoma (NHL), and multiple myeloma (MM). An oligo-based microarray platform was employed with higher density coverage for the DNA repair/ repair-related genes (1 probe/0.2–7.0 kb). Results were assessed relative to multiple databases of known benign variation. A proof of principal analysis included RB1, TP53 and ATM, for which genomic alterations and disease preference have been well established. Analysis of RB1 revealed CNAs in 74 cases (4 gains and 70 losses) (17.4% of total) with copy gains uniformly attributable to trisomy 13 in cases of B-ALL. As expected, loss of RB1 typically reflected either monosomy 13 (31/70 seen predominately in MM) or smaller deletions (∼1 Mb, seen in CLL). CNAs in TP53 included 2 gains and 42 losses (10.4% of total), with gains again related to trisomy (exclusively hyperdiploid B-ALL), whereas deletions varied in size and disease association. CNAs in the ATM included 28 gains and 19 losses (11.1% of total). Gains reflecting trisomy correlated with diverse diagnoses, while more focal gains (∼10 Mb) were seen exclusively in B-ALL (3 cases). Not surprisingly, small losses of ATM (<10 Mb) occurred almost exclusively in CLL. Thus, microarray findings for these classic tumor suppressor genes correlated well with disease specificity for alterations submicroscopic in size (<10 Mb).

An additional 152 DNA repair/repair-related genes were analyzed and a number of genes showed disease-specific correlations for submicroscopic CNAs. For example, deletion of ERCC5 was seen in B-ALL (2 cases); however, co-deletion of ERCC5 and LIG4 was confined to NHL (3 cases). Smaller deletions involving a number of other genes also occurred specifically in NHL (GTF2H4, POLQ, REV1L, FANCL, BRCA2, and NUDT1). While abnormalities in TERT were infrequent (21 gains and 4 losses) and nearly always associated with aneuploidy, an intragenic duplication of three exons was seen in a single case of diffuse large B-cell lymphoma with a history of failed therapy. Regarding AML, deletion of UBE2V2 was seen in a case exhibiting numerous multi-megabase amplifications, a potential consequence of dysregulated DNA replication. AML-specific CNAs include deletions of RPA2 that always associated with translocation-positive cases (PML/RARA; MLL/MLLT3; MLL/MLLT4), and deletions of APTX (reduced expression correlates with favorable therapy response), as well as an intronic duplication within MGMT (therapy resistance associates with overexpression). Smaller CNAs involving ALKBH3 were seen specifically in MM (3 cases), while deletions of ALKBH5 impacting nearby TOP3A were seen in MDS (2 cases). Additional genes in which recurrent CNAs were associated with multiple disorders included NEIL1, PMS2, ENDOV, CHEK2, and HELQ.

The data reveal CNAs in DNA repair and repair related genes; many are recurrent abnormalities that exhibit disease specificity. The findings are consistent with previous data addressing biological function, disease progression and response to therapy. Further study is warranted to validate these alterations as diagnostic/prognostic biomarkers or as predictive biomarkers useful in directing more personalized treatment decisions. They may also represent potential new targets for therapy.