The Preclinical Efficacy of a Novel Telomerase Inhibitor, Imetelstat, in AML - a Randomized Trial in Patient-Derived Xenografts

Replicative immortality depends on telomerase activation in the majority of cancers including acute myeloid leukemia (AML). Imetelstat is a covalently lipidated 13-mer oligonucleotide that competitively inhibits telomerase activity. Clinical efficacy of imetelstat has recently been reported in essential thrombocythemia and myelofibrosis. We investigated the efficacy of imetelstat in AML using a randomized trial in patient-derived xenografts (PDX).

To establish an AML PDX cohort, primary bone marrow (BM) or peripheral blood (PB) samples from 31 AML patients were transplanted into NOD.Cg Prkdc^scid Il2rg^tm1Wjl Tg (CMV-IL3,CSF2,KITLG)1Eav/Mlo (NSGS). Engraftment was defined by reconstitution of BM and spleen with CD33+ donor cells, PB circulating blasts, anemia or thrombocytopenia. The success rate for engraftment was 70.4% and was independent of cytogenetics.

Fifteen independent AML patient samples (n=12 recipients/sample) were tested for efficacy of imetelstat. After engraftment, mice were randomized and treated with imetelstat (15 mg/kg body weight) or control phosphate-buffered saline (PBS) intraperitoneally every 48-72h. Across the entire cohort, survival was improved for imetelstat- vs. PBS-treated PDX (Hazard ratio PBS: 5.299; 3.379 ± 8.312; p < 0.0001; Cox proportional hazards model). Imetelstat delayed the expansion of human AML cells in 14 PDX (93.3%). AML patient samples were divided into 2 groups based on survival outcomes, "Sustained responders" (imetelstat extended survival > 2-fold or > 4 weeks; 9 AML patient samples; 60%) or "Poor responders" (6 AML patient samples, 40%). European LeukemiaNet (ELN)-karyotypic subtypes differed trend-wise between groups: favourable (100% sustained, 0% poor response to imetelstat); intermediate-1, intermediate-2 and poor (45.5% sustained, 54.5% poor response to imetelstat; p = 0.1, Fisher's exact test). Next generation sequencing revealed baseline AML PDX genetic profiles. FLT3, NRAS, TET2, RAD21 as well as 15 genes annotated in DNA damage, cell cycle regulation and apoptosis were mutated exclusively in the sustained responders group. Poor responders revealed mutations in ETV6, FANCM, MKI67, WT1 and 16 additional genes regulating growth factor independence and evasion of apoptosis. Imetelstat response was associated with reduced quiescence and induction of γ-H2AX levels in AML populations (n=4-6 / group / 3 individual AML patient samples): γ-H2AX (MFI normalized to PBS; PBS: 1.0 ± 0.02; imetelstat: 1.24 ± 0.02; p < 0.0001, Student's t test), quiescence (%Ki67-diploid AML cells; PBS: 11.62% ± 0.87; imetelstat: 6.9% ± 0.68; p < 0.001, Student's t test). The presence of spliceosome mutations (SRSF2, U2AF1) was not correlated with response (p = 0.5; Fisher's exact test).

To investigate the effects on normal human hematopoiesis, CD34-enriched cord blood was transplanted into NOD.Cg-Prkdc^scidIl2rg^tm1Wjl/SzJ (NSG, n=5-6/group/donor in 2 replicate experiments). Engraftment was confirmed at 4 weeks post-transplant by PB chimerism (>=1%). Mice were treated with imetelstat or vehicle control from 4 weeks. After 10 weeks treatment, BM cellularity and donor engraftment were significantly reduced whereas the frequency of the CD34+CD38low HSC-enriched population was unchanged (donor 1 PBS: 0.35% ± 0.03%, imetelstat: 0.23% ± 0.06%, p = 0.1; donor 2 PBS: 0.45% ± 0.05%, imetelstat: 0.92% ± 0.35%, p = 0.2, Student's t test). HSC cell cycle and γ-H2AX levels were unchanged. PB and spleen B-cell engraftment were reduced (donor 1 spleen PBS: 77.74% ± 4.66%, imetelstat: 45.48% ± 9.03%, p < 0.05; donor 2 spleen PBS: 93.95% ± 0.40%, imetelstat: 89.1% ± 2.52%, p = 0.09, Student's t test), whereas the myeloid lineage was elevated or relatively preserved (donor 1 spleen PBS: 3.29% ± 0.71%, imetelstat: 7.53% ± 1.6%, p < 0.05; donor 2 spleen PBS: 1.20% ± 0.12%, imetelstat: 4.49% ± 1.60%, p = 0.07, Student's t test).

In summary, imetelstat demonstrates efficacy in a significant proportion of AML PDX (60%, 9 out of 15 individual AML patient samples). Robust responses to imetelstat were associated with favourable cytogenetic risk groups and mutations in pathways controlling DNA damage. The effects on normal human hematopoiesis were modest and predominantly seen in the B-lymphocyte lineage with relative preservation of myeloid and stem cell populations. Further study is warranted to understand the preclinical and clinical efficacy of imetelstat in AML.