Fighting AML with its own weapons

In this issue of Blood, Mill et al¹ report that acquired loss-of-function mutations of RUNX1 impair ribosomal biogenesis, rendering myeloid progenitor cells in acute myeloid leukemia (AML) susceptible to apoptotic elimination following combined inhibition of protein translation and BCL-2 (see figure). Harnessing AML-specific adaptations in the protein translation machinery for therapeutic exploitation to enhance susceptibility to apoptosis has recently been tested preclinically.² This study demonstrates that for patients with RUNX1-mutant AML this approach allows for a more efficient impairment of leukemic cell growth.

RUNX1 (AML1) is a DNA sequence-specific transcriptional master regulator governing gene expression programs of hematopoietic development and differentiation. Its encoding gene is subject to recurrent germline or somatic hotspot mutations that are associated with several hematologic malignancies and leukemia predisposition syndromes. Underscoring their clinical uniqueness, patients bearing RUNX1 lesions are separate entities in the standing World Health Organization classification of AML.

Broadly, two classes of gene alterations have been identified with distinct molecular and hematologic disease phenotypes: monoallelic chromosomal translocations and mono- or biallelic somatic mutations.³ Translocations, such as t(8;21)(q22;q22) and t(12;21)(p13;q22), largely retain RUNX1’s DNA-binding capacity and exert neomorphic activities, as opposed to fusion lesions, which have a favorable prognosis, deletions, missense, splicing, frameshift, and nonsense mutations have a loss-of-function or dominant negative activity. The latter mutations are associated with poor prognosis,³ which may be due to the resistance of RUNX1-mutant (mutRUNX1) AML to mainstay chemotherapies, albeit mechanistic insights have been sparse.

To elucidate clinically relevant molecular consequences of RUNX1 loss-of-function mutations, Mill et al developed a set of preclinical models of human AML, including AML cell lines, patient-derived primary cells, as well as 2 novel isogenic AML cell lines bearing homozygous or heterozygous RUNX1^R174* mutations. The authors show that homozygous and heterozygous RUNX1^R174*-mutant AML cells had compromised myeloid differentiation potential at the morphological and molecular levels compared with RUNX1 competent controls (see figure). This is consistent with the known role of RUNX1 in regulating transcription of and in concert with key hematopoietic transcription factors through chromatin remodeling activity and enhancer activation to drive hematopoietic differentiation.^4,5

Interestingly, monoallelic RUNX1^R174* was sufficient to also decrease c-MYC and perturb c-MYC dependent gene expression. c-MYC is a global regulator of transcription, chromatin structure, DNA replication, and protein synthesis.⁶ Accordingly, the authors found decreased total RNA levels and molecular perturbations consistent with reduced ribosomal biogenesis and protein translation in monoallelic RUNX1^R174*-mutant AML cells. Protein synthesis is tightly regulated and governed by a complex machinery; modulation of protein translation rates can confer unique cellular properties to stem cells⁷ and therapy resistance in cancer.⁸ An elegant preclinical study from Speck and colleagues has uncovered substantially increased resistance of Runx1-deficient mouse nonleukemic hematopoietic stem and multipotent progenitor cells to genotoxic stress-induced ablation. The investigators attributed this resistance to reduced protein translation rates and altered susceptibility to apoptosis induction.⁸ 

Probing whether reduced ribosomal biogenesis may render mutRUNX1 AML cells vulnerable to further inhibition of protein translation using homoharringtonine (HTT), a natural cytotoxic alkaloid from the evergreen tree known to interfere with chain elongation, or its analog omacetaxine (OM), the authors discovered more severely compromised protein synthesis at various levels in mutant compared with RUNX1 wild-type AML cells. This was accompanied by reduced levels of short-lived proteins, including cMYC and antiapoptotic MCL1 and BCL-xL. Mechanistically, the authors attributed the effects of HTT to epigenetic inactivation and impaired expression of MYC targets, specifically genes encoding transfer RNAs and others facilitating RNA translation in mutRUNX AML cells.

Notably, key myeloid differentiation instructing regulators, such as RUNX1 itself and one of its most critical targets, PU.1,⁴ were also reduced at the protein level in HTT-treated RUNX1^R174*/WT AML cells compared with wild-type counterparts. Despite this, HTT-exposed mutRUNX1 AML cells displayed morphological and molecular signs of myeloid maturation and decreased viability, indicating the restoration of differentiation programs either independent of, or otherwise suppressed by the two master regulators. Whether and how much of the gene regulatory differences observed in treatment-naïve and HTT-exposed mutRUNX1 AML cells is attributable to loss of RUNX1 alone, a combination of impaired RUNX1 and PU.1 protein dosage, or potentially perturbed recruitment dynamics⁵ of these master regulators and cooperating lineage-specific transcription factors is an emerging question with relevance to potentially extending this approach to targeting aberrant myeloid cell growth in the absence of genetic RUNX1 lesions.

The authors further explored whether reduced levels of key antiapoptotic proteins, MCL1 and BCL-xL, found in HTT-treated mutRUNX1 AML cells, increased the cells’ susceptibility to apoptotic cell death. And indeed, impairment of protein synthesis using HTT or OM rendered AML cells more susceptible to pharmacologic BCL-2 inhibition by venetoclax or BRD2/3/4 BET inhibitor (OTX015) (known to inhibit expression of apoptosis inhibitors). HTT/OM synergized with venetoclax and BET inhibition in reducing cell viability of mutRUNX1 and wild-type AML cells, while only moderately compromising nonleukemic hematopoietic stem and progenitor cells ex vivo, and curtailing mutRUNX1 AML cell growth in patient-derived xenotransplantation models in vivo.

Insights gained from this work provide a rationale for the clinical application of combined protein translation and BCL-2 inhibitors in mutRUNX1 AML. This paper also poses critical follow-up questions. Perhaps the most pressing being: Will this combination treatment allow for the elimination of (pre) leukemic stem cells and reduce the risk of not eliminating all disease-relevant cancer cells that drive AML relapse and progression? Although restoration of apoptosis signaling may not be sufficient for the eradication of leukemic stem cells,⁹ it is possible that inhibition of protein synthesis, especially mitochondrial protein translation,¹⁰ may curtail metabolic adaptations driving venetoclax resistance of leukemic stem cells⁹ and allow for the effective eradication of all malignant cells in patients. Further investigations following this important work will undoubtedly continue leading our way for harnessing AML’s defenses as our offense.