Background
TP53 mutant AMLs are a clinically distinctive group of AMLs often associated with poor clinical outcome and resistance to conventional chemotherapy.1 While infrequent in de novo AMLs, they are more frequent in older patients and therapy related AMLs.1 TP53 mutations are also associated with venetoclax (VEN) resistance, which is often used to treat older patients ineligible for more intensive conventional therapy.2 3 Thus, there is a need to identify alternative therapeutic approaches to treat TP53 mutant AMLs, especially in the context of VEN resistance.
Methods
Differential expression comparing VEN resistant and sensitive AML cell-lines was used to derive a VEN resistant signatures. Non-negative matrix factorization (NMF) was used to decompose this signature in the BeatAML4 cohort and identify distinct transcriptional VEN resistant clusters/states (VR_C1-4).3 Frequency of TP53 mutations was assessed in these states. Gene/protein expression, drug response and gene essentiality data from BeatAML, TCGA5 and CCLE6 were used to characterize transcriptionally distinct groups of TP53 mutant AMLs. We also evaluated drug response in vitro in AML cell-lines.
Results
TP53 mutations predominantly occurred in VR_C2 and 3, with sporadic occurrence in other states across cohorts (figure 1A). TP53 mutants in VR_C3 (VR_C3@MUTs) were transcriptionally distinct from other TP53 mutants (not_C3@MUTs) in both patients and cell-lines (figure 1B). VR_C3@MUT were characterized by over-expression of several inflammatory, metabolic and DNA damage response pathways (figure 1C). VR_C3@MUT also had activated JAK-STAT signaling (figure 1C-D), which translated to specific sensitivity to pharmacological and genetic inhibition of JAK pathway (figure 2A-B). We confirmed this sensitivity in cell-lines models using the JAK inhibitor ruxolitinib (RUXO), which intriguingly in a subset of VR_C3@MUTs RUXO improved response to VEN (figure 2C), providing a rationale for combination therapy. In addition to high expression of inflammatory pathways, VR_C3@MUTs were also characterized by increased infiltration by cytotoxic T lymphocytes (CTLs) and high expression of and checkpoint genes (figure 3). These data suggest that this subset of TP53 mutant AMLs may benefit from immune checkpoint therapy (ICT).
Conclusions
We identify a transcriptionally distinct groups of VEN resistant TP53 mutant AMLs that are sensitive to JAK inhibition. They also have high CTL infiltration and expression of immune checkpoint genes, presenting a rationale to explore ICT alone or in combination with JAK inhibitors and VEN as a therapeutic strategy in these AMLs.
References
Daver NG, et al. TP53-Mutated myelodysplastic syndrome and acute myeloid leukemia: biology, current therapy, and future directions. Cancer Discov, 2022;12(11):2516–2529.
Pollyea DA, et al. Venetoclax with azacitidine or decitabine in patients with newly diagnosed acute myeloid leukemia: Long term follow-up from a phase 1b study. Am J Hematol, 2021;96(2):208–217.
Mohanty V, et al. Transcriptional and phenotypic heterogeneity underpinning venetoclax resistance in AML. bioRxiv, 2024.
Bottomly D, et al. Integrative analysis of drug response and clinical outcome in acute myeloid leukemia. Cancer Cell, 2022;40(8):850–864 e9.
Salmon JM, et al. Epigenetic activation of plasmacytoid DCs Drives IFNAR-dependent therapeutic differentiation of AML. Cancer Discov, 2022;12(6):1560–1579.
Barretina J, et al. The cancer cell line encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature, 2012;483(7391):603–7.
Abstract 1394 Figure 1
A) TP53 mutants (in red) across cohorts. B) PCA plot showing transcriptionally distinct sub-groups in TP53 mutants. C) Pathways differentially expressed in VR_C3@MUT relative to other groups D) JAK activity across groups inferred from RPPA
Abstract 1394 Figure 2
Boxplots depicting sensitivity to JAK inhibitors in BeatAML (A) and essentiality of JAK2 in CCLE (B) across groups. C) Heatmap of area under the drug response curve (AUC) of AML cell-lines treated with VEN, RUXO and their combination
Abstract 1394 Figure 3
A subset of TP53 mutant AMLs are highly infiltrated. A) Abundance of CTLs inferred from bulk gene expression. B) Heatmap of fold-changes of immune checkpoint genes in VR_C3@MUTs relative to other groups