Aimee Paterson

Understanding the genetic and nongenetic determinants of tumor protein 53 (TP53)-mutation-driven clonal evolution and subsequent transformation is a crucial step toward the design of rational therapeutic strategies. Here we carry out allelic resolution single-cell multi-omic analysis of hematopoietic stem/progenitor cells (HSPCs) from patients with a myeloproliferative neoplasm who transform to TP53-mutant secondary acute myeloid leukemia (sAML). All patients showed dominant TP53 'multihit' HSPC clones at transformation, with a leukemia stem cell transcriptional signature strongly predictive of adverse outcomes in independent cohorts, across both TP53-mutant and wild-type (WT) AML. Through analysis of serial samples, antecedent TP53-heterozygous clones and in vivo perturbations, we demonstrate a hitherto unrecognized effect of chronic inflammation, which suppressed TP53 WT HSPCs while enhancing the fitness advantage of TP53-mutant cells and promoted genetic evolution. Our findings will facilitate the development of risk-stratification, early detection and treatment strategies for TP53-mutant leukemia, and are of broad relevance to other cancer types.

Summary TP53 is the most commonly mutated gene in human cancer, typically occurring in association with complex cytogenetics and dismal outcomes. Understanding the genetic and non-genetic determinants of TP53- mutation driven clonal evolution and subsequent transformation is a crucial step towards the design of rational therapeutic strategies. Here, we carry out allelic resolution single-cell multi-omic analysis of haematopoietic stem/progenitor cells (HSPC) from patients with a myeloproliferative neoplasm who transform to TP53- mutant secondary acute myeloid leukaemia (AML), a tractable model of TP53 -mutant cancer evolution. All patients showed dominant TP53 ‘ multi-hit’ HSPC clones at transformation, with a leukaemia stem cell transcriptional signature strongly predictive of adverse outcome in independent cohorts, across both TP53- mutant and wild-type AML. Through analysis of serial samples and antecedent TP53 -heterozygous clones, we demonstrate a hitherto unrecognised effect of chronic inflammation, which supressed TP53 wild-type HSPC whilst enhancing the fitness advantage of TP53 mutant cells. Our findings will facilitate the development of risk-stratification, early detection and treatment strategies for TP53 -mutant leukaemia, and are of broader relevance to other cancer types.

Abstract In myeloid malignancies, presence of 'multi-hit' TP53 mutations is associated with lack of response to conventional therapy and dismal outcomes, particularly when found in combination with a complex karyotype. Therefore, it is crucial to understand the biological basis of TP53-mutant driven clonal evolution, suppression of antecedent clones and eventual disease transformation to inform the development of more effective therapies. Myeloproliferative neoplasms (MPN) represent an ideal tractable disease model to study this process, as progression to secondary acute myeloid leukemia (sAML) frequently occurs through the acquisition of TP53 missense mutations. To characterize tumor phylogenies, cellular hierarchies and molecular features of TP53-driven transformation, we performed single-cell multi-omic TARGET-seq analysis (PMID: 33377019 & 30765193) of 22116 hematopoietic stem and progenitor cells (HSPCs) from 35 donors and 40 timepoints (controls, MPN in chronic phase, pre-AML and TP53-mutated sAML; Figure1a). TARGET-seq uniquely enables single-cell mutation analysis with allelic resolution with parallel transcriptomic and cell-surface proteomic readouts. We invariably identified convergent clonal evolution leading to complete loss of TP53 wild-type alleles upon transformation, including parallel evolution of separate TP53 "multi-hit" subclones in the same patient (n=4/14) and JAK2-negative progression (n=2/14). Complex clonal evolution driven by chromosomal abnormalities (CAs) was present in all patients and TP53 multi-hit HSPCs without CAs were rarely observed. Subclones with recurrent CA such as monosomy 7 showed upregulation of RAS-associated transcription and preferentially expanded in xenograft models. Together, these data indicate that TP53 missense mutation, loss of TP53 wild-type allele and cytogenetic evolution are collectively required for leukemic stem cell (LSC) expansion. Integrated transcriptomic analysis of sAML samples (Figure1b) revealed three major populations: (1) a TP53-mutant cluster (Figure1c) characterized by an erythroid signature (e.g. KLF1, GATA1, GYPA; an unexpected finding as no cases showed diagnostic features of erythroid leukemia), (2) an LSC TP53-mutant cluster (Figure1d) and (3) a TP53-WT preleukemic cluster (Figure1e). The LSC cluster showed dysregulation of key stem cell regulators, from which we derived a novel 48-gene LSC score with prognostic impact in an independent AML cohort (HR=3.13; Figure1f). Importantly, this score was predictive of outcome irrespective of TP53 status for both de novo and sAML, demonstrating its broader potential clinical utility. TARGET-seq analysis uniquely allowed us to characterize rare TP53-WT preleukemic cells (preLSCs), which were almost exclusively confined to the immunophenotypic lineage-CD34+CD38-CD90+CD45RA- HSC compartment. PreLSC from sAML samples presented increased stemness, increased quiescence, aberrant inflammatory signaling and differentiation defects (Figure1g) as compared to HSCs from normal or MPN donors, both at the transcriptional and functional levels through in vitro long-term and short-term cultures. This indicates cell-extrinsic suppression of residual TP53-WT hematopoiesis. Longitudinal analysis of TP53-heterozygous mutant HSPCs at different stages of disease evolution (Figure1a) revealed that aberrant inflammatory signalling (e.g. BST2, IFITM1, IFITM3) in the genetic ancestors of TP53 "multi-hit" LSCs, but not the presence of TP53-mutations alone, was predictive of subsequent transformation. In a mouse model system, TP53-mutant cells challenged with sustained inflammatory stimuli acquired a mean 3-fold competitive advantage in WT: TP53 R172H/+chimeras. This indicates that pro-inflammatory cues from the tumour microenvironment promote fitness advantage of TP53-mutant cells whilst supressing antecedent clones. In summary, we present a comprehensive single-cell multi-omic analysis of the genetic, cellular and molecular landscape of TP53-mediated transformation, providing unique insights into the evolution of chronic hematological malignancies towards an aggressive acute leukemia (Figure1h). Since TP53 is the most commonly mutated gene in human cancer, we anticipate these findings will be of broader relevance to many other cancer types. Figure 1 Figure 1. Disclosures Kretzschmar: Vanadis Diagnostics, a PerkinElmer company.: Current Employment. Drummond: BMS: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; CTI: Membership on an entity's Board of Directors or advisory committees; Novartis: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau. Harrison: Geron: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; BMS: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Galacteo: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Keros: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Sierra Oncology: Honoraria; Constellation Pharmaceuticals: Research Funding; Abbvie: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; AOP Orphan Pharmaceuticals: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Incyte Corporation: Speakers Bureau; Promedior: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Janssen: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Roche: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Shire: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Gilead Sciences: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; CTI BioPharma: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Celgene: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding, Speakers Bureau; Novartis: Membership on an entity's Board of Directors or advisory committees, Research Funding, Speakers Bureau. Mead: Abbvie: Consultancy, Honoraria; Celgene/BMS: Consultancy, Honoraria, Research Funding; Novartis: Consultancy, Honoraria, Speakers Bureau.

Myeloproliferative neoplasms (MPN) are heterogenous clonal hematologic neoplasms where current therapies show limited disease-modification. We previously reported that the splicing factor, SF3B1, is mutated in 5-10% of MPN correlating with myelofibrotic progression in essential thrombocythemia. This adverse phenotype contrasts with SF3B1 mutationsin myelodysplasia where it is associated with milder disease. Moreover, SF3B1 K666 is the dominant hotspot in MPN in contrast to K700E in MDS (abstract #185043). The mechanism by which SF3B1 mutations accelerate myelofibrotic progression in JAK2V617F-mutated MPN is not understood and will be a key step towards the development of disease-modifying therapies. To address this, we applied two complimentary single-cell multiomic methods to study 15 individuals (healthy donor [HD, n=5], JAK2V617F-single mutated [J+, n=5] and JAK2V617F- SF3B1K666double mutated MPN cases [JS+, n=5]). First, we characterised the cellular landscape of JS+ MPN using CITE-seq/10X genomics platform. Then, we analyzed SF3B1-mutant ( SF3B1mut) disease-driving hematopoietic stem and progenitor cells (HSPCs) and their associated aberrant splicing signatures using TARGET-seq, which permitted single-cell genotyping with allelic resolution and intra-patient comparison of mutant versus wild-type (wt) cells. In an analysis of 109,498 cells, we observed abnormally expanded erythroid (ERP), megakaryocyte (MK) and unexpectedly, eosinophil-basophil-mast cell (EBM) progenitors in JS+ as compared with J+ or HD HSPCs (Fig 1A). Two HSC clusters were present, one of which (HSC_mpp2) principally comprised of JS+ cells and had a distinct molecular program (Fig 1A); upregulation in MK (eg. PF4, VWF) and EBM lineage signature genes (eg. TPSB2), interferon signaling genes but downregulation in apoptosis genes and canonical stemness markers (eg. CD133). These signatures suggest the JS+ driven HSC cluster may be MK/EBM-transcriptionally primed and may therefore, promote expansion of these abnormal progenitors which are implicated in driving fibrosis development. Through genotyping of >5000 HSPCs, we resolved the clonal architecture of JS+ MPN, observing that in most cases double mutant cells were present and clonally dominant, with mutant SF3B1 the initiating event. Mutations arose from the HSC level and genotypes were equally distributed in the HSPC compartment. Highlighting the importance of single-cell analysis, and not apparent at the bulk level, JAK2V617F and SF3B1K666mutationsresided in separate clones for one case with a WHO fibrosis score 0 in contrast to the other 4 cases with double mutant cells and fibrosis scores 1-3. This reinforces the cooperative effect of thesemutationsto drive fibrosis. We performed full-length transcriptomics of select HSPCs to study SF3B1mut-specific aberrant splicing events (ASE). In a global splicing analysis, more events were detected in JAK2V617F- SF3B1K666 double mutant than single mutant HSPCs. Retained intron (RI) events were the most prevalent ASE, followed by skipped exon and alternative 3' splice site events. We identified previously described SF3B1mut-specific ASE ( MAP3K7, ERGIC3 and SEPTIN6) confirming the validity of our dataset to study new ASE. Inter- and intra-patient comparison of SF3B1mut versus SF3B1wt cells uncovered a number of novel biologically relevant ASE, including two RI events in STAT1. These RI events were reduced in SF3B1mut HSPCs with the predicted consequence of increased STAT1 expression and activation, of direct relevance to MPN pathobiology since STAT1 is known to play a key role in enhancing megakaryopoiesis in MPN (Fig 1B). We validated these events using long-read Nanopore sequencing of single-cell cDNA libraries and have developed a cell model system with transient expression of SF3B1K666 mutation versus SF3B1wt in HEK293T to corroborate STAT1 ASE occurring as a direct consequence of mutant SF3B1. In further support of upregulated STAT1 activation, we observed upregulation of STAT1 downstream targets in abnormally expanded JS+ HSPCs. In conclusion, this analysis resolved the clonal architecture and hematopoietic cellular composition of JS+ MPN, identifying abnormally expanded and transcriptionally primed HSPC populations with abnormal splicing of STAT1 enhancing JAK-STAT signaling, uncovering the mechanism by which these mutations promote an accelerated fibrotic phenotype.