Transcriptional variability accelerates preleukemia by cell diversification and perturbation of protein synthesis

Abstract

Transcriptional variability facilitates stochastic cell diversification and can in turn underpin adaptation to stress or injury. We hypothesize that it may analogously facilitate progression of premalignancy to cancer. To investigate this, we initiated preleukemia in mouse cells with enhanced transcriptional variability due to conditional disruption of the histone lysine acetyltransferase gene Kat2a . By combining single-cell RNA sequencing of preleukemia with functional analysis of transformation, we show that Kat2a loss results in global variegation of cell identity and accumulation of preleukemic cells. Leukemia progression is subsequently facilitated by destabilization of ribosome biogenesis and protein synthesis, which confer a transient transformation advantage. The contribution of transcriptional variability to early cancer evolution reflects a generic role in promoting cell fate transitions, which, in the case of well-adapted malignancies, contrastingly differentiates and depletes cancer stem cells. That is, transcriptional variability confers forward momentum to cell fate systems, with differential multistage impact throughout cancer evolution.

Contributions of transcriptional noise to leukaemia evolution: KAT2A as a case-study

“…Nevertheless, the scenarios are not mutually exclusive. The two pre-leukaemia models studied by us in Gupta et al [106] suggest specific dependencies or interactions between the mutational events and the gene expression programmes that respond to KAT2A loss with enhanced noise. Transcriptional noise is not sufficient to initiate the pre-leukaemia state, as Kat2a NULL mice do not develop myeloproliferation or indeed leukaemia [96,105].…”

“…In a case-study, we used Kat2a loss by Mx1-Cre recombinase-dependent genetic knockout to generate Kat2a NULL haematopoietic stem cells and progenitors [ 105 ] in which AML was initiated through retroviral delivery of fusion oncogenes [ 105 , 106 ] or concomitant recombination of an oncogenic allele (e.g. Idh1 R132H ) [ 106 ].…”

“…In a case-study, we used Kat2a loss by Mx1-Cre recombinase-dependent genetic knockout to generate Kat2a NULL haematopoietic stem cells and progenitors [ 105 ] in which AML was initiated through retroviral delivery of fusion oncogenes [ 105 , 106 ] or concomitant recombination of an oncogenic allele (e.g. Idh1 R132H ) [ 106 ]. As observed in mES cells, loss of Kat2a increased transcriptional noise [ 105 ], more markedly so in genes whose promoters lost H3K9ac, supporting a role for histone acetylation in preserving stable locus activity [ 107 ].…”

“…As observed in mES cells, loss of Kat2a increased transcriptional noise [ 105 ], more markedly so in genes whose promoters lost H3K9ac, supporting a role for histone acetylation in preserving stable locus activity [ 107 ]. Downstream of the enhanced transcriptional noise, Kat2a NULL cells explored a wider transcriptional space, with enhanced cellular heterogeneity and increased cell differentiation outputs [ 105 , 106 ]. The basic observations of molecular and cellular diversification upon Kat2a knockout are similar to the consequences of KAT2A inhibition in mES cells [ 89 ], but the effects on leukaemia evolution are not uniform ( figure 3 ).…”

“…Indeed, it has been modelled that persistent transcriptional noise is incompatible with terminal cell differentiation [ 110 ], a conclusion supported by the Kat2a NULL data. Cellular diversification is also observed in Kat2a NULL pre-leukaemia states [ 106 ], but the consequences for disease progression and survival outcome are distinct. In this case, Kat2a NULL cells carrying the pre-leukaemic genetic event gained an advantage that allowed them to progress through the process of leukaemia transformation more efficiently than wild-type (WT) pre-leukaemia cells, resulting in higher AML penetrance and poorer survival.…”