Cohesin Haploinsufficiency Drives Leukemic Initiation through Fli1 Upregulation in Inv(16) AML

Acute myeloid leukemia (AML) is a genetically diverse cancer of the bone marrow that affects approximately 20,000 new patients annually in the US alone. Many mutations drive leukemogenesis, including those in the core binding factor (CFB) complex, which is a heterodimeric transcription factor involved in hematopoietic cell development and differentiation (Pulikkan and Castilla 2018). Mutations in the CBF complex are present in 15% of adult AML cases and can affect either the alpha (RUNX1) or the beta (CBFβ) subunit through the chromosomal aberrations t(8;21) and inv(16), respectively. These mutations are necessary but not sufficient for leukemic initiation.

CBF mutations co-occur with other mutations, including those affecting the cohesin complex, which is mutated in 10% of all AML cases (Ley et al 2013). The cohesin complex is comprised of four subunits, RAD21, SMC1A, SMC3, and STAG1/2, and regulates chromatin architecture and genomic looping, thereby affecting transcription. Mutations in the cohesin complex in AML affect only one subunit, resulting in haploinsufficiency of the complex and loss of function (Heimbruch et al. 2021).

Despite CBF leukemias being classified as one group clinically, t(8;21) and inv(16) have different co-mutational profiles (Faber et al. 2016, Opatz et al. 2020, Duployez et al. 2016, Jahn et al. 2020, Qin et al. 2022). Interestingly, cohesin complex mutations have been observed more commonly in t(8;21) AML than in inv(16) AML. We therefore hypothesized that haploinsufficiency of cohesin would be detrimental to the development or maintenance of inv(16) AML. To test this, we utilized the inv(16) model published by Kuo et al. 2006 and the Smc3 haploinsufficiency model published by Viny et al. 2015. Both inv(16) and the loss of a single Smc3 allele are driven by MX1-Cre activation by PIPC injection.

In contrast to our hypothesis, we found that inv(16);Smc3^+/- mice developed AML with a shorter latency (14 weeks) than did inv(16);Smc3^+/+ mice (20 weeks), suggesting the two mutations cooperate during leukemic initiation. ATAC sequencing of HSPCs revealed increased global accessibility in inv(16);Smc3^+/- vs. inv(16);Smc3^+/+ cells, particularly at Ets family transcription factor binding sites. qPCR and western blotting confirmed an upregulation the Ets family member Fli1 in inv(16);Smc3^+/- cells. Upregulation of several known Fli1 target genes such as Ccng1, Ccl2, Mdm2, and Rab27b, was also observed. Interestingly, upregulation of Fli1 was not observed in leukemic cells from inv(16);Smc3^+/- mice, suggesting cohesin mutation and Fli1 upregulation may contribute to leukemic initiation but not maintenance. In support of this, knockdown of Rad21 induced apoptosis in leukemias from inv(16);Smc3^+/+ mice. We next performed single cell RNA sequencing on lin-, Sca1+, c-Kit+ (LSK) sorted cells from both genotypes to identify which cell types exhibited elevated Fli1. We found a significant upregulation of Fli1 in several clusters, primarily in LSKs, ST-HSCs, and MPPs, suggesting that Fli1 upregulation occurs in the most primitive stem cells. We are currently testing if Fli1 upregulation drives a shortened latency using inv(16);Smc3^+/+ HSPCs.

Cohesin mutations have not been readily detected in inv(16) AML, with current data primarily including patients of Chinese and German/Austrian origin (Faber et al. 2016, Opatz et al. 2020, Duployez et al. 2016, Jahn et al. 2020, Qin et al. 2022). Studies that have detected cohesin mutations in inv(16) patients have used a targeted sequencing approach in addition to whole genome or exome sequencing. Collectively, this suggests that cohesin mutations may co-occur with inv(16) at a higher rate than reported, which needs to be confirmed with targeted sequencing of a more diverse population. Another possibility is that cohesin mutations primarily affect the initiation of inv(16) AML but have a negative impact on maintenance. Therefore, selective pressure may no longer exist at the AML stage, leading to mutational loss. It would be interesting to examine the clonal evolution of healthy persons harboring inv(16) mutations in an effort to detect gain or loss of cohesin mutations. Our data support the hypothesis that cohesin mutations drive inv(16) AML initiation through the upregulation of Fli1 and its downstream targets in HSPCs and underline the importance of studying the effect of mutations on both AML initiation and maintenance.

No relevant conflicts of interest to declare.