Juvenile myelomonocytic leukemia (JMML) is a myeloproliferative neoplasm of young children that originates from early hematopoietic stem/progenitor cells. We have previously developed an in vivo disease model using xenotransplantation of primary JMML cells into Rag2-/-γc-/- mice (Haematologica 2016;101:597). The model reproduces a characteristic JMML phenotype including myelomonocytic proliferation, hepatosplenomegaly, and lung infiltration. Case-specific driver mutations and DNA methylation patterns are unchanged after xenologous engraftment, indicating their origin in leukemia-initiating cells.

We and others recently discovered a tight link between prognosis and differential DNA methylation in JMML (Nat Commun 2017;8:2126; Nat Commun 2017;8:2127; Blood 2018;131:1576). We also reported that the DNA methyltransferase inhibitor 5-azacytidine (azacitidine, 5AC) has unprecedented clinical activity in JMML and induces complete or partial remissions before allogeneic HSCT (Blood 2015;125:2311). Cytosine arabinoside (araC) is a structurally related nucleoside which is commonly used for cytoreduction in JMML but lacks the ability to induce remissions. Here we employed the xenotransplantation model to investigate the antileukemic activity and epigenetic effects of 5AC on JMML in comparison to araC.

After eight weeks of leukemic expansion, 15 xenograft mice were treated with two cycles of 5AC (3 mg/kg/d x 5 days every two weeks). Control groups included mice treated with araC 20 mg/kg/d (N=15) or carrier solution (0.9% NaCl, N=20). The experimental animals maintained stable body weight, and no major toxicity on murine hematopoiesis was observed. 5AC and araC exhibited antileukemic activity and substantially reduced the human JMML cell content in bone marrow, spleen, liver, and lung. However, we noted that CD34+ stem/progenitor cells within the human leukemia population were depleted after treatment with 5AC but not after araC (5AC, 20.2% +/- 7.3%; araC, 35.6% +/- 6.1 %; carrier, 39.4% +/- 3.5%; p<0.01). To demonstrate that the selective reduction of CD34+ cells impaired the leukemia-initiating capacity of the xenograft, we treated a subsequent series of mice as above and retransplanted the bone marrow into secondary recipient mice. JMML cells obtained from 5AC-treated primary recipients sustained engraftment in only one of 9 secondary recipients at 30 weeks after retransplantation whereas JMML xenografts treated with araC or carrier engrafted in 8/13 or 4/8 secondary mice, respectively (p=0.03).

We then examined the genome-wide DNA methylation in 5AC-treated xenografts (N=5) using Infinium 450K arrays. The JMML genomes exhibited global and profound DNA demethylation with near-complete loss of fully methylated CpG sites. A focused analysis of approximately 5,000 CpG sites with JMML-specific methylation illustrated that the profiles of 5AC-treated JMML cells were more similar to healthy human CD34+ cells than untreated JMML cells. As expected, no change in DNA methylation was observed in xenografts treated with araC.

Next we studied the early effects of 5AC on the transcriptome and epigenome of JMML. Xenograft mice were treated as above, and JMML cells were harvested from bone marrow on days 0, 2, 4, and 6. RNA sequencing readily identified non-random changes in gene transcription that progressed over time from days 2 to 6 and were reproducible across replicate mice. Between days 0 and 6 we observed >2fold upregulation of 856 transcripts and downregulation of 958 transcripts (<0.01 false discovery rate, multitest-corrected). CpG-rich 5' regions (putative promoters) of corresponding genes were invariably demethylated. Gene Ontology enrichment analysis linked upregulated transcripts to myeloid differentiation whereas downregulated transcripts were involved in nucleosome assembly/organization and chromatin silencing.

In summary, the xenograft experiments highlight the therapeutic potential of 5AC in JMML and thus encourage the further clinical development of epigenetic therapy with hypomethylating agents for this disease.


Niemeyer:Celgene: Consultancy, Membership on an entity's Board of Directors or advisory committees.

Author notes


Asterisk with author names denotes non-ASH members.