Acute Myeloid Leukemia (AML) Patients with an Isolated Trisomy Display Increased Frequency of Spliceosome Mutations

Introduction: The molecular landscape of patients (pts) with cytogenetically normal (CN) AML has been extensively evaluated; however, there are limited data providing comprehensive mutation analysis of AML pts with sole somatic trisomy. A few analyses with small numbers of pts found frequent mutations of spliceosome encoding genes (Eisfeld et al, Leukemia 2016, Herold et al, Blood 2014).

Methods: Review of the Cancer and Leukemia Group B/Alliance for Clinical Trials in Oncology cytogenetics database found 138 de novo adult AML pts with isolated trisomy of chromosomes 4, 8, 11, 13 or 21. We compared clinical and molecular data between pts within each isolated trisomy group as well as with 131 pts who had ≥1 chromosome abnormality in addition to a trisomy (non-sole trisomy) and with 716 pts with CN-AML.

Results: Spliceosome encoding genes, mainly SRSF2 and U2AF1, were often mutated in pts with isolated trisomy (Figure 1). SRSF2 was mutated in 28% of +8 pts and 35% of +21 pts. SRSF2 mutations were the most common mutation in +13 pts, seen in 48% of +13 pts. U2AF1 was mutated in 15% of +8 pts and in 43% of +11 pts. Other genes commonly mutated in each trisomy group included DNMT3A, FLT3 -ITD, IDH2, NRAS, RUNX1 and TET2. Notably, pts with +4 lacked mutations in ASXL1, BCOR, STAG2 and WT1 genes, all of which were mutated in pts in the other sole trisomy groups. NPM1 mutations, often seen in the +4, +8, +13 and +21 groups, were not seen in +11 pts. There were no differences in complete remission rates, relapse rates, disease-free or overall survival among trisomy groups. For pre-treatment clinical features of the 138 de novo isolated trisomy AML pts see Table 1. The median age was 61 years (range, 19-84 years).

We saw several differences in the mutation spectrum between older (aged ≥60 years) and younger (aged <60 years) sole +8 pts. Older +8 pts had more mutations in spliceosome encoding genes (69% vs 18%, P <.001) and transcription factors (55% vs 9%, P <.001), and less mutations in genes affecting protein kinase function (21% vs 60%, P=.004). Older +8 pts (n=30) had more mutations in SRSF2 (50% vs 6%, P <.001), RUNX1 (34% vs 9%, P=.02) and TET2 (25% vs 6%, P=.04), and tended to have less FLT3 -ITD (P=.09) and NPM1 (P=.08) mutations than younger (n=32) +8 pts. Older +8 pts also had a higher median number of mutations than their younger counterparts (P=.05).

We next studied molecular differences between the 138 sole trisomy pts and 131 pts with non-sole trisomy. We saw more mutations in SRSF2 (29% vs 14%, P=.003) and U2AF1 (14% vs 14%, P=.003) in sole trisomy versus non-sole trisomy pts. Several genes were often mutated (≥5%) in both sole and non-sole trisomy groups; most common were spliceosome genes SRSF2 , U2AF1, ZRSR2, and others were ASXL1, BCOR, DNMT3A, FLT3 -ITD, FLT3 -TKD, IDH1, IDH2, NPM1, NRAS, RUNX1, STAG2, TET2. TP53 mutations were often seen in non-sole trisomy pts, but were rare in sole trisomy pts (17% vs 1%, P <.001). Compared to pts with non-sole trisomy, sole trisomy pts presented with more mutations (median, 3 vs 2, P <.001), most affecting kinase function (37% vs 18%, P=.001), DNA methylation (64% vs 39%, P <.001), RNA splicing (48% vs 24%, P <.001) and transcription (36% vs 23%, P=.02), while pts with non-sole trisomy had more mutations of the RAS pathway (27% vs 15%, P=.02) and tumor suppressor function (24% vs 8%, P <.001). Clinically, pts with non-sole trisomy were younger (medians: 58 vs 61 years, P=.07) and had higher pre-treatment leukocyte counts (medians: 14.6 vs 9.6 x10⁹/L, P=.09) and more blood (medians: 48% vs 37%, P=.03) and marrow (medians: 72% vs 65%, P=.05) blasts than pts with sole trisomy. Also, non-sole trisomy pts more often presented with extramedullary disease (26% vs 14%, P=.03).

Comparing sole trisomy pts to 716 CN-AML pts, we found spliceosome encoding genes more often mutated (48% vs 19%, P <.001). We saw more mutations in CEBPA (P=.01), DNMT3A (P=.08), FLT3 -ITD (P=.02), and NPM1 (P <.001 ) in CN-AML pts.

Conclusions: This is the largest series of de novo AML pts with isolated trisomy that provides comprehensive mutation analysis. These pts display a distinct molecular landscape from AML pts with non-sole trisomy and from CN-AML pts. Mutations affecting spliceosome encoding genes were seen in almost 50% of pts with isolated trisomy suggesting that aberrant RNA splicing may be implicated in the pathogenesis of these pts.

Support: U10CA180821, U10CA180850, U10CA180882, U24CA196171; ClinicalTrials.gov Ids: NCT00048958, NCT00900224

View largeDownload slide

View largeDownload slide

Close modal