Inhibition of DOCK1 prevents the clonal expansion of high-risk TP53-mutant clonal hematopoiesis induced by genotoxic stressors. Free

Background: TP53-mutant clonal hematopoiesis (CH) is prevalent in populations that received genotoxic cancer therapy and is associated with elevated risk for therapy-related myeloid neoplasms. As TP53 mutations primarily act through loss of function, it has been difficult to develop TP53-mutant-specific pharmacologic inhibitors. The concept of synthetic lethality (SL) presents a promising strategy for targeting TP53-mutant CH. Here, we developed a human model of TP53-mutant CH that exhibits clonal expansion upon exposure to genotoxic stressors and identified a targetable SL interaction that may prevent this clonal expansion.

Methods: To model human TP53-mutant CH clonal expansion under genotoxic treatments, we disrupted the TP53 locus (or AAVS1 “safe harbor” locus as control) in primary human hematopoietic stem and progenitor cells (HSPCs) using CRISPR/Cas9. We then screened various cancer therapies and selected olaparib and paclitaxel for promotion of TP53-mutant CH. Finally, we employed a computational algorithm – Mining Synthetic Lethals (MiSL) – to identify candidate SL partners for mutant TP53, and functionally investigated candidates in our human TP53-mutant CH model.

Results: We disrupted the TP53 gene in primary human HSPCs through CRISPR/Cas9, reflected by a 100% variant allele frequency (VAF) of Insertions/Deletions (InDels) at the TP53 locus (TP53^KO). In a 1:1 co-culture with AAVS1 HSPCs over 3 weeks, olaparib- or paclitaxel-treated TP53^KO HSPCs exhibited an increased competitive advantage compared to the untreated control. Specifically, the VAF of TP53^KO increased from 19.2% in the untreated group to 59% and 94% after paclitaxel or olaparib treatment, respectively. In vivo and in the absence of genotoxic agents, engraftment of TP53^KO HSPCs was comparable to that of AAVS1 HSPCs. However, after 1mg/kg paclitaxel treatment in vivo, human chimerism in mice engrafted with TP53^KO HSPCs significantly increased from 5.7% to 48.3% (p=0.01, n=4), while no increase in engraftment was found in mice engrafted with AAVS1 HSPCs. To facilitate tracking of clonal expansion by flow cytometry, we generated TP53^mut-GFP⁺ HSPCs using CRISPR/AAV6-mediated homology-directed repair. Similarly to previous experiments, olaparib treatment selected for TP53^mut-GFP⁺cells, whereas vehicle-treated cultures showed no selection for TP53^mut-GFP⁺cells in a 1:1 co-culture assay (average 94.8% versus 43.8% GFP+, p=0.0146).

Using this system, we aimed to gain insight into the mechanisms of clonal selection.TP53^KO and AAVS1 HSPCs showed comparable IC50 to olaparib (176µM vs. 319µM) and paclitaxel (7.3nM vs 21nM), suggesting that mechanisms underlying the competitive advantage of TP53-mutant HSPC may not be differential apoptosis of bulk HSPCs. We then assessed whether TP53^KO impacts hematopoietic stem cell (HSC) frequency and self-renewal specifically. We performed limiting dilution transplantation assays after pretreatment with olaparib. Here, we found that olaparib-treated TP53^KO HSPCs had a higher frequency of functional HSCs compared to olaparib-treated AAVS1 HSPCs (p=0.016, Chi square test). Similarly, in serial replating colony forming unit assays, olaparib- or paclitaxel-pretreated TP53^KO HSPCs produced 4.1-fold or 4.7-fold more colonies compared to controls in the first (p= 0.002 & 0.001), 13.5-fold and 16.6-fold in the second (p= 0.05 & 0.006), and 8.4-fold and 17.4-fold in the third plating (p=0.008 & 0.008), respectively. These observations indicated increased HSC frequency and enhanced self-renewal capacity in TP53^KO HSPCs compared to AAVS1 after genotoxic treatment.

Finally, we used the MiSL algorithm to identify DOCK1 as a potential SL partner based on the mutual exclusivity of DOCK1 disruptions with TP53-mutations in AML (p=0.0009). Pharmacological inhibition of DOCK1 effectively abolished the clonal expansion of TP53^mut-GFP⁺ HSPCs in vitro under olaparib treatment. In a 1:1 co-culture, the normalized percentage of TP53^mut-GFP⁺ cells after 3 weeks was 50.0% with the vehicle control, 43% with DOCK1i (p=0.27), 82% with olaparib (p=0.01), and 54% with simultaneous olaparib and DOCK1 inhibitor (p=0.02).

Conclusion: Our TP53-mutant human CH models faithfully recapitulates the clonal expansion of TP53-mutant CH upon exposure to genotoxic stressors, including olaparib and paclitaxel. Further, we identify DOCK1 as a potential SL partner with mutant TP53 demonstrating that inhibition of DOCK1 can prevent clonal expansion of TP53^mut CH.