Telomere Content Diversity in Myeloid Neoplasia Could Inform on Differential Sensitivity to Telomerase Inhibitors

The recent FDA approval of Imetelstat, a new class of antineoplastic drugs for the treatment of MDS1 opens the question on whether telomere content (TC) and/or telomerase function might constitute therapeutic targets and diagnostic biomarkers. Early studies have not specifically delved in elucidating such possibility. In general, cancer relies on excessive shortening of telomeres, and one could envision that systematic and comprehensive genomic and functional studies of TC in myeloid neoplasia (MN) might reveal the “Achilles' heel” e.g., the group with a TC likely responding to Imetelstat.

We have adopted WGS-based pipelines to measure TC in MN. Taking advantage of a large, well-annotated cohort (n=1804) of patients with MN including AML (n=730), MDS (n=702), MPN (n=372) and 163 non-malignant controls (PNH/AA [n=102], persistent polyclonal B lymphocytosis (n=50) and healthy subjects [n=11]) we studied TC in correlation with pathomorphologic, cytogenetic and molecular subtypes of the disease. First, we cross-validated different bioinformatic WGS-based pipelines23 and established consistency with PCR-based TC measurements. Thus, we observed a general decrease in TC in MN patients irrespective of age. In both AML and MDS cohorts (for subsequent analysis referred together as MN), blast percentage correlated with telomere shortening (p<.0001, r²=0.0231 and p=.0346, r²=0.0063, respectively). Comparing TC among karyotypic abnormalities in MN, we detected higher TC in patients with del(5q) (p<.0001) and lower TC in patients with t(6;9) (p=.0002), t(8;21) (p<.0001), inv(16) (p=.0003), t(9;11) (p<.0001), t(15;17) (p<.0001) when compared to patients with normal karyotype.

The analysis of TC according to the genetic landscape revealed higher TC in patients harboring ETV6 (p=.0172), SF3B1 (p=.0015), SRSF2 (p=.0005) and TP53 (p=.0025) mutations and lower TC in patients carrying FLT3 (p<.0001), JAK2 (p=.0191), KRAS (p=.0001), NPM1 (p<.0001), NRAS (p=.0234) and WT1 (p=.0002) mutations when compared to the overall MN population.

We then explored specific TC signatures and we observed that high TC MN (above 90th percentile) were mainly represented by MDS and were enriched in -7/del(7q), inv(3)/t(3;3) and complex karyotype abnormalities as well as ASXL1, SRSF2 and TP53 mutations. Low TC MN were mainly represented by AML and were enriched in t(15;17) and t(8;21) as well as KRAS, NRAS, NPM1 and WT1 mutations.

In a multivariate analysis nor the diagnosis (AML vs MDS), the age or the karyotype (abnormal vs normal) correlated to TC, whereas we detected an independent role of blast percentage (p<.0001) and TP53 mutations (positively correlating with TC, p=.001).

To corroborate our findings, we investigated germline mutations in genes involved in telomere machinery (18 variants identified, 2 co-occurring with TP53 mutation [ATRX, c.6332G>C; TERT, c.1807del]) and excluded an impact of them on TC in this context. RNASeq supported these findings as mRNA expression of genes involved in telomere maintenance (RAP1A, TERC, TINF2, TPP1, CTC1) were only increased in patients with MN and high TC.

Functional studies, including singletons composition of telomeres investigation, C-circle analysis and transcriptome-based scores to assess telomerase activity (EXTENDScore), showed that among all mutations, the ones with profound genomic alterations (e.g., TP53 mutations) maintain telomere elongation activity.

In conclusion, our findings demonstrate that telomere shortening is common in MN irrespective of telomere machinery hypomorphic variants. In addition, mutations driving hyperproliferation (e.g., NRAS, KRAS) might exceed the compensatory capacity of normal telomere lengthening as also suggested by the negative correlation between TL and blast percentage.

These results open the question as to whether TC might be used as a marker of response to Imetelstat and whether can identify disease phenotypes more likely to respond to the drug (MN relying on high TC (e.g., TP53 mutants) vs MN reaching a critical level of telomeres shortening (e.g., NRAS/KRAS mutated).