Microdeletions at Apparently Balanced Translocation Breakpoints Are Biomarkers for Heightened Genomic Instability in Patients with Non-MLL Acute Myeloid Leukemia

Abstract 2529

The clinical heterogeneity of acute myeloid leukemia (AML) is well established and identification of specific balanced translocations is a critical component of the classification of AML. Patients with translocations are assigned to prognostic groups generally defined as good [e.g., t(8;21)(q22;q22), inv(16) (p13q22), t(15;17)(q24;q21)], intermediate [e.g., t(3;5)(q25;q34), t(9;11)(p22;q23)] and poor [e.g., other MLL translocations, inv(3)(q21q26.2), t(6;9)(p23;q34)]. Further refinement based on the identification of new genetic markers may help tailor prognostic and treatment decisions. In an effort to identify new molecular markers associated with recurrent translocations, we have analyzed 30 cases of AML with apparently balanced translocations using a microarray-based technology coupled with linear DNA amplification [translocation-CGH (tCGH)], providing high resolution mapping of translocation breakpoints and identification of copy number abnormalities at those breakpoints (BkPt-CNAs). Analysis can be multiplexed to interrogate 14 balanced translocations in a single test. Additionally, all cases were assayed by high-density aCGH to identify additional clinically relevant copy number abnormalities (CNAs) throughout the genome. For each sample these test results were compared. Patients were all adults (>40 years of age), 10 males and 20 females, and included t(15;17)(q24;q21) (8 cases), t(8;21) (q22;q22) (6 cases), inv(16)(p13q22) (5 cases), t(9;11)(p22;q23) (5 cases), t(3;5)(q25;q34) (4 cases), t(6;11)(q27;q23) (1 case) and t(6;9)(p23;q34) (1 case).

Of the 30 cases, 6 showed BkPt-CNAs in one or both translocation partners. All 6 (100%) of these cases exhibited additional clinically relevant CNAs, whereas only 11/24 (46%) cases with no BkPt-CNAs showed additional CNAs. Thus, BkPt-CNAs are significantly correlated with the presence of secondary clinically relevant alterations (p=0.0208). Importantly, no MLL translocations cases revealed BkPt-CNAs. In contrast, 3 of 4 t(3;5)(q25;q34) cases showed BkPt-CNAs in both genes (NPM1 and MLF1) and additional CNAs, with two cases showing large deletions (>0.5 Mb) at both breakpoints. Alterations involved RSRC1 (mRNA splicing), FBXW11 (E3 ubiquitin protein ligase), STK10 (serine/threonine kinase), FGFR3 (growth factor receptor) and MGMT (O-6-methylguanine-DNA methyltransferase), findings potentially relevant to disease progression and treatment response. Cases of t(15;17), t(8;21) and inv(16) each showed a single case with BkPt-CNAs and these correlated with greater complexity for additional CNAs. For t(15;17), that case showed partial duplication of both the PML and RARA genes (47 kb and 39 kb) yielding a cryptic fusion (missed by karyotype and FISH) that produced the “long” form (bcr1) transcript based on the PML breakpoint and confirmed by rtPCR. This case also had a 24.56 Mb deletion at 13q14.2->q21.33 involving RB1, MIR15A and MIR16-1. In the 7 remaining PML-RARA cases [4 short (bcr3) and 3 long (bcr1)] with no BkPt-CNAs, two cases showed non-complex CNAs [(e.g., -Y (bcr1) or +8 (bcr3)]. Five of six t(8;21) cases exhibited CNAs, but the case with a BkPt-CNA (39 kb within RUNX1) exhibited greater complexity for CNAs (multiple gains and losses over >70 MB of distal 9q) than other cases (e.g., +8 or +15). Similarly, analysis of inv(16) cases demonstrated CNAs in 3 of 5 cases, with the greatest complexity in the case with BkPt-CNAs [deletion in CBFB and MYH11 involving ABCC1 (anion transporter implicated in multiple drug resistance), deletions 9q34.13 involving ABL1 and NUP214], while other cases had single changes of a homozygous deletion of GSTT1 or a heterozygous deletion of the 16p11.2 microdeletion syndrome region, either potentially inherited. The sole case of t(6;9) had only the relevant translocation.

In conclusion, analysis of AML cases with balanced translocation by t-CGH and aCGH revealed significant correlation between BkPt-CNAs and the presence of secondary CNAs. Moreover, when BkPt-CNAs were present, the molecular complexity of additional CNAs was increased. This observation held true across various translocations evaluated, with the notable exception of MLL translocations. Thus BkPt-CNAs provide a broadly applicable signature for a subtle molecular instability in AML that, acquired or inherited, may predate translocations and may prove relevant to treatment response and outcome.