Normal hematopoiesis is organized in a hierarchical manner and it has been hypothesized that acute myeloid leukemia (AML) is organized in a similar way with leukemia-initiating cells (LIC) at the top of the hierarchy, giving rise to more differentiated blasts to sustain AML. Therefore, elimination of LIC population is critical for cure. This may be accomplished via novel molecular targeted therapies. The mutational composition of LIC and non-LIC compartments in AML has not been fully elucidated and could provide new insights into biology and treatment. We investigated the distribution and variant allelic frequencies (VAFs) of recurrent gene mutations within these compartments in newly diagnosed CD34+ AML patients (pts).
We studied a total of 88 pts. CD34- AML cases, defined as <5% positivity on blasts, were excluded. Pre-treatment bone marrow or apheresis samples were sorted and sequenced with our ultrasensitive limited cell (LC)-FACS-seq method. First, we gated on CD45dimLin- leukemic population, followed by isolation of 300 cells from CD34+CD38- (LIC), CD34+CD38+ (non-LIC) and CD34- compartments. To compare with the bulk population, DNA was extracted from 500,000 CD45dimLin- leukemic blasts. All samples were sequenced with a 27-gene targeted panel. Extreme Limited Dilution Analysis (ELDA) platform was used for colony formation assays and estimation of stem cell frequencies.
Clinical characteristics are summarized in Table 1. The median frequency of the LIC population was 0.5% (range, 0.01% - 69%). The prevalence of high LIC frequency (≥0.5%) was significantly higher in pts with adverse risk (AR) AML, as compared to intermediate (IR) and favorable risk groups (94% vs 34% vs 16%, respectively, p<.001). When compared to pts with low LIC frequency (<0.5%), those with high LIC frequency had worse overall survival (median, 9 months vs not reached, p=.003) and relapse-free survival (median, 4 vs 15 months, p=.01). In 10 pts who had serial relapse samples, LIC frequencies were increased at the time of relapse (p=.03).
We re-validated the commonly used LIC markers with ELDA of primary AML cells. In one IR sample, stem cell frequencies in sorted CD34+CD38-, CD34+CD38+ and CD34- compartments were 1:3, 1:15 and 1:16, respectively (p<.001). In one AR sample, stem cell frequencies were 1:1, 1:8, and 1:12, respectively (p<.001). Using these markers, LICs and non-LICs were enriched and sequenced. The average number of mutations detected by sequencing of bulk samples was significantly lower than sorted LIC (3.17 vs 3.75, p<.05) and non-LIC (3.17 vs 3.96, p<.001) populations indicating the higher sensitivity of our method in detecting subclonal mutations. Mean VAFs were similar between LIC and non-LIC populations for NPM1 (42% vs 47%), DNMT3A (37% vs 41%), IDH1 (41% vs 48%), IDH2 (43% vs 48%), and U2AF1 (37% vs 42%) mutations. Mutations involving signaling pathways were more frequent in non-LICs, including FLT3-TKD (12% vs 23%, p<.01), NRAS (17% vs 26%, p<.01) and KRAS (13% vs 19%, p<.05) mutations, which might be explained by their later acquisition during AML development. In addition, among 22 pts with CEBPA mutation, 16 (73%) harbored the mutation exclusively in non-LICs. Finally, 13 pts with TP53 mutations had different VAFs between compartments. Among 4 pts who had doubling of VAF from LIC to non-LIC compartment, 3 had subclones with del(17p) in LIC pool detected by FISH. LIC subclones harboring both del(17p) and TP53 mutation (i.e. loss of heterozygosity) propagated to drive leukemia.
Relapse samples obtained from 6 pts were analyzed and compared with diagnosis. In all cases, we could identify LIC clones that persisted after chemotherapy and led to relapse (see example in Figure). Similarly, 3 pts who were primary refractory showed persistence of LIC clones that were resistant to treatment. On the contrary, 6 pts in whom LIC clones could be eradicated with treatment did not experience disease recurrence.
LICs exist at a very low frequency in pre-treatment AML samples. The mutational composition of LIC-enriched compartment shows differences from blasts constituting the bulk of leukemia, which is consistent with the sequence of mutations observed during the evolution of AML. LC-FACS-seq is an ultrasensitive method to detect mutations in a tiny population of residual LICs in pts at remission. Therapies targeting mutations that are concentrated in LICs may re-shape the clonal hierarchy and impact on disease course.
Behbehani:Fluidigm corporation: Other: Travel funding. Byrd:Ohio State University: Patents & Royalties: OSU-2S; Genentech: Research Funding; Genentech: Research Funding; Janssen: Consultancy, Other: Travel Expenses, Research Funding, Speakers Bureau; Janssen: Consultancy, Other: Travel Expenses, Research Funding, Speakers Bureau; BeiGene: Research Funding; BeiGene: Research Funding; Gilead: Other: Travel Expenses, Research Funding, Speakers Bureau; BeiGene: Research Funding; Gilead: Other: Travel Expenses, Research Funding, Speakers Bureau; Novartis: Other: Travel Expenses, Speakers Bureau; Gilead: Other: Travel Expenses, Research Funding, Speakers Bureau; Novartis: Other: Travel Expenses, Speakers Bureau; Pharmacyclics LLC, an AbbVie Company: Other: Travel Expenses, Research Funding, Speakers Bureau; Novartis: Other: Travel Expenses, Speakers Bureau; Pharmacyclics LLC, an AbbVie Company: Other: Travel Expenses, Research Funding, Speakers Bureau; TG Therapeutics: Other: Travel Expenses, Research Funding, Speakers Bureau; Genentech: Research Funding; Pharmacyclics LLC, an AbbVie Company: Other: Travel Expenses, Research Funding, Speakers Bureau; Acerta: Research Funding; TG Therapeutics: Other: Travel Expenses, Research Funding, Speakers Bureau; Acerta: Research Funding; Janssen: Consultancy, Other: Travel Expenses, Research Funding, Speakers Bureau; Ohio State University: Patents & Royalties: OSU-2S; Ohio State University: Patents & Royalties: OSU-2S; Acerta: Research Funding; TG Therapeutics: Other: Travel Expenses, Research Funding, Speakers Bureau. Lozanski:Boehringer Ingelheim: Research Funding; Beckman Coulter: Research Funding; Stemline Therapeutics Inc.: Research Funding; Genentec: Research Funding.
Author notes
Asterisk with author names denotes non-ASH members.